research scientific journal articles

Trending Articles

• The Fanconi anemia pathway induces chromothripsis and ecDNA-driven cancer drug resistance. Engel JL, et al. Cell. 2024. PMID: 39181133
• Impact of preexisting interstitial lung disease on outcomes of lung cancer surgery: A monocentric retrospective study. Goga A, et al. Respir Med Res. 2024. PMID: 39182360
• Phase III KEYNOTE-789 Study of Pemetrexed and Platinum With or Without Pembrolizumab for Tyrosine Kinase Inhibitor‒Resistant, EGFR -Mutant, Metastatic Nonsquamous Non-Small Cell Lung Cancer. Yang JC, et al. J Clin Oncol. 2024. PMID: 39173098
• Advances and future directions in ROS1 fusion-positive lung cancer. Boulanger MC, et al. Oncologist. 2024. PMID: 39177972
• A line attractor encoding a persistent internal state requires neuropeptide signaling. Mountoufaris G, et al. Cell. 2024. PMID: 39191257

Latest Literature

• Clin Infect Dis (3)
• J Clin Endocrinol Metab (1)
• J Clin Invest (1)
• J Natl Cancer Inst (1)
• Neurology (3)
• PLoS One (63)
• Sci Transl Med (6)

NCBI Literature Resources

MeSH PMC Bookshelf Disclaimer

The PubMed wordmark and PubMed logo are registered trademarks of the U.S. Department of Health and Human Services (HHS). Unauthorized use of these marks is strictly prohibited.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

PubMed Central (PMC) Home Page

PubMed Central ® (PMC) is a free full-text archive of biomedical and life sciences journal literature at the U.S. National Institutes of Health's National Library of Medicine (NIH/NLM)

Discover a digital archive of scholarly articles, spanning centuries of scientific research.

Learn how to find and read articles of interest to you.

Collections

Browse the PMC Journal List or learn about some of PMC's unique collections.

For Authors

Navigate the PMC submission methods to comply with a funder mandate, expand access, and ensure preservation.

For Publishers

Learn about deposit options for journals and publishers and the PMC selection process.

For Developers

Find tools for bulk download, text mining, and other machine analysis.

10.2 MILLION articles are archived in PMC.

Content provided in part by:, full participation journals.

Journals deposit the complete contents of each issue or volume.

NIH Portfolio Journals

Journals deposit all NIH-funded articles as defined by the NIH Public Access Policy.

Selective Deposit Programs

Publisher deposits a subset of articles from a collection of journals.

March 21, 2024

Preview upcoming improvements to pmc.

We are pleased to announce the availability of a preview of improvements planned for the PMC website. These…

Dec. 15, 2023

Update on pubreader format.

The PubReader format was added to PMC in 2012 to make it easier to read full text articles on tablet, mobile, and oth…

We are pleased to announce the availability of a preview of improvements planned for the PMC website. These improvements will become the default in October 2024.

Navigation group

Home banner.

Where scientists empower society

Creating solutions for healthy lives on a healthy planet.

9.4 million

2.8 billion

article views and downloads

Main Content

• Editors and reviewers
• Collaborators

Find a journal

We have a home for your research. Our community led journals cover more than 1,500 academic disciplines and are some of the largest and most cited in their fields.

Submit your research

Start your submission and get more impact for your research by publishing with us.

Author guidelines

Ready to publish? Check our author guidelines for everything you need to know about submitting, from choosing a journal and section to preparing your manuscript.

Peer review

Our efficient collaborative peer review means you’ll get a decision on your manuscript in an average of 61 days.

Article publishing charges (APCs) apply to articles that are accepted for publication by our external and independent editorial boards

Press office

Visit our press office for key media contact information, as well as Frontiers’ media kit, including our embargo policy, logos, key facts, leadership bios, and imagery.

Institutional partnerships

Join more than 555 institutions around the world already benefiting from an institutional membership with Frontiers, including CERN, Max Planck Society, and the University of Oxford.

Publishing partnerships

Partner with Frontiers and make your society’s transition to open access a reality with our custom-built platform and publishing expertise.

Policy Labs

Connecting experts from business, science, and policy to strengthen the dialogue between scientific research and informed policymaking.

How we publish

All Frontiers journals are community-run and fully open access, so every research article we publish is immediately and permanently free to read.

Editor guidelines

Reviewing a manuscript? See our guidelines for everything you need to know about our peer review process.

Become an editor

Apply to join an editorial board and collaborate with an international team of carefully selected independent researchers.

My assignments

It’s easy to find and track your editorial assignments with our platform, 'My Frontiers' – saving you time to spend on your own research.

Three-step plan to cut overlooked methane emissions could help us stop global warming faster

Methane, the second most important greenhouse gas, has been neglected—but now scientists lay out a new plan and a new tool to cut emissions down to size

Safeguarding peer review to ensure quality at scale

Making scientific research open has never been more important. But for research to be trusted, it must be of the highest quality. Facing an industry-wide rise in fraudulent science, Frontiers has increased its focus on safeguarding quality.

Ecopreneur Prof Thomas Crowther to showcase the power of nature-based solutions at Frontiers Forum virtual event

Visionary ecologist and pioneer in reforestation and ecosystem restoration Prof Thomas Crowther will explore the vital role that nature plays in our fight against climate change, at a unique Frontiers Forum virtual event on 11 September.

Pompeii skeleton discovery shows another natural disaster may have made Vesuvius eruption even more deadly

The death of inhabitants of Pompeii during the eruption of Vesuvius in 79CE is mostly attributed to volcanic causes, such as falling ashes and hot gas. This, however, may not be the full picture.

Dune-inspired upgrade for spacesuits allow astronauts to recycle urine into water

Inspired by the sci-fi series, Dune, Cornell researchers created a prototype novel urine collection and filtration system for spacesuits.

What happens when the ocean’s biggest predators start hunting each other? Here are five Frontiers articles you won’t want to miss

At Frontiers, we bring some of the world’s best research to a global audience. But with tens of thousands of articles published each year, it’s impossible to cover all of them. Here are just five amazing papers you may have missed.

Five Research Topics on the human impact of population growth

We highlight five Research Topics tackling some of the most urgent issues emerging from our planet's growing population.

Get the latest research updates, subscribe to our newsletter

Quick Links

• RESOURCE CENTER
• MEMBER LOGIN

External Links

• AAAS Communities
• SCIENCE CAREERS
• SCIENCE FAMILY OF JOURNALS
• More AAAS Sites

Scientific Journals

The 2024 Noto earthquake, with a magnitude of about 7.5, resulted in a large amount of ground shaking across the Noto Peninsula and Central Japan. Consequences included damaged houses in the Shinmachi district, Suzu, Ishikawa, Japan. The fault system responsible for the earthquake ruptured in a complex manner, generating a devastating event. See pages 866 and 871. Photo: The Asahi Shimbun via Getty Images

German research vessel Maria S. Merian leaving port for the northwest African coast. Sediment gravity flows are agents of sediment transport and deposition, but these enormous events can pose risks to seafloor infrastructure. Böttner et al. used field data from the Agadir Canyon off the coast of Morrocco to map the deposits and erosive destruction of a giant submarine gravity flow from source to sink. They found that entrainment of mud promoted extreme erosion and run-away flow growth. Credit: Sven Heinrich

Specialized Macrophages Repair the Lung After Damage. Airborne pathogens and pollutants can cause long-lasting lung damage, but how lung myeloid cells contribute to postinjury repair is not fully understood. Ruscitti et al. identified a population of Ly6G+ macrophages that emerges during the early recovery phase after influenza virus–triggered lung injury in mice. These Ly6G+ macrophages were short-lived, monocyte-derived cells that populated alveoli near the site of damage and promoted epithelial repair. This month’s cover illustration depicts damaged lung alveoli as burning buildings, from which falling debris (representing signals such as GM-CSF and type 2 cytokines) promotes monocyte differentiation into reparative Ly6G+ macrophages. Credit: Adeline Deward/Illumine

A Variable-Stiffness Deformable Wheel. Wheels have a low cost of transport over flat ground but struggle to overcome large obstacles. Inspired by the surface tension of a water droplet, Lee et al. developed a morphing wheel that rolls over flat ground in the circular high-modulus state and deforms over obstacles in the low-modulus state. The modulus of the wheel is changed in real time by adjusting the tension in the spoke structure. This month’s cover depicts a two-wheeled mobile platform with variable-stiffness wheels in the low-modulus state rolling over an irregular-shaped rock outdoors. Credit: Sinabro Studio

This week, Mahmud et al. report that human cytomegalovirus (HCMV) ensures its persistence by preventing the full activation of the kinase Akt in monocytes. The image shows HCMV captured by transmission electron microscopy. Credit: James Cavallini/Science Source

A Breath of Fresh Air. The cover shows a scanning electron microscopy image of pulmonary blood vessels from a mouse after surgical lung resection. A potential consequence of lung transplantation or injury is impaired regeneration of lung vasculature and intussusceptive angiogenesis, or split blood vessels. To promote regeneration over fibrotic injury, Ma et al. used a pneumonectomy mouse model of lung injury where a portion of the lung was removed. The authors found that genetically silencing Ras homolog family member J (Rhoj), which is expressed by a subset of endothelial cells, could improve lung function and decrease the number of split blood vessels. They linked the improvement to mRNA modification of the Forkhead box protein O1 gene transcript in endothelial cells. This approach also improved lung function in a bleomycin-induced fibrosis model. These data suggest that targeting endothelial Rhoj could be a new avenue for treating lung injury. Credit: Ma et al./Science Translational Medicine

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts

Research articles

Microwaves induced epitaxial growth of urchin like MIL-53(Al) crystals on ceramic supports

• Limor Ben Neon
• Martin Drobek

‘The mirror of the soul?’ Inferring sadness in the eyes

• Jonas Moosavi
• Annika Resch
• Marina A. Pavlova

Association between change in heart rate over years and life span in the Paris Prospective 1, the Whitehall 1, and Framingham studies

• Eugenie Valentin
• Xavier Jouven

Developing an image-based grading scale for peripheral drusen to investigate associations of peripheral drusen type with age-related macular degeneration

• Paripoorna Sharma
• Fritz Gerald P. Kalaw
• Shyamanga Borooah

Physics-Informed Masked Autoencoder for active sparse imaging

• Luke McEvoy
• Daniel Tafone
• Yuping Huang

The synergistic effect of triglyceride-glucose index and HbA1c on blood pressure control in patients with hypertension: a retrospective cohort study

• Jinghan Hai
• Xiaofeng Li

A meta-analysis of performance advantages on athletes in multiple object tracking tasks

• Hui Juan Liu

Prediction model of weight control experience in men with obesity in their 30 s and 40 s using decision tree analysis

• Myeunghee Han

Various distance between generalized Diophantine fuzzy sets using multiple criteria decision making and their real life applications

• Murugan Palanikumar
• Nasreen Kausar
• Fikadu Tesgera Tolasa

Preparation and performance control of ultra-low near-infrared reflectivity coatings with super-hydrophobic and outstanding mechanical properties

• Weigang Zhang
• Yueting Zhuang
• Qianfeng Zhang

Adaptive condition-aware high-dimensional decoupling remote sensing image object detection algorithm

• Chenshuai Bai
• Xiaofeng Bai

Functional prediction of response to therapy prior to therapeutic intervention is associated with improved survival in patients with high-grade glioma

• Aubrey Ledford
• Analiz Rodriguez
• Teresa M. DesRochers

The epidemiology of lung cancer in Hungary based on the characteristics of patients diagnosed in 2018

• István Kenessey
• Petra Parrag
• Csaba Polgár

Evaluation of newly synthesized 2-(thiophen-2-yl)-1H-indole derivatives as anticancer agents against HCT-116 cell proliferation via cell cycle arrest and down regulation of miR-25

• Nagwa M. Abdelazeem
• Shaimaa A. Gouhar
• Marwa El-Hussieny

Bobcat Optimization Algorithm: an effective bio-inspired metaheuristic algorithm for solving supply chain optimization problems

• Zoubida Benmamoun
• Khaoula Khlie
• Youness Gherabi

CAPZB mRNA is a novel biomarker for cervical high-grade squamous lesions

• Wanqiu Huang

Analysis of oceanic suspended particulate matter in the western North Pacific using the complex amplitude sensor

• Atsushi Yoshida
• Yutaka Tobo
• Makoto Koike

Polypeptide N -acetylgalactosaminyltransferase-15 regulates adipogenesis in human SGBS cells

• Asuka Takahashi
• Takeo Iwata

Effect of breed and dietary composition on the miRNA profile of beef steers divergent for feed efficiency

• D. A. Kenny

Temporal witnesses of non-classicality in a macroscopic biological system

• Giuseppe Di Pietra
• Vlatko Vedral
• Chiara Marletto

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Explore millions of high-quality primary sources and images from around the world, including artworks, maps, photographs, and more.

Explore migration issues through a variety of media types

• Part of The Streets are Talking: Public Forms of Creative Expression from Around the World
• Part of The Journal of Economic Perspectives, Vol. 34, No. 1 (Winter 2020)
• Part of Cato Institute (Aug. 3, 2021)
• Part of University of California Press
• Part of Open: Smithsonian National Museum of African American History & Culture
• Part of Indiana Journal of Global Legal Studies, Vol. 19, No. 1 (Winter 2012)
• Part of R Street Institute (Nov. 1, 2020)
• Part of Leuven University Press
• Part of UN Secretary-General Papers: Ban Ki-moon (2007-2016)
• Part of Perspectives on Terrorism, Vol. 12, No. 4 (August 2018)
• Part of Leveraging Lives: Serbia and Illegal Tunisian Migration to Europe, Carnegie Endowment for International Peace (Mar. 1, 2023)
• Part of UCL Press

Harness the power of visual materials—explore more than 3 million images now on JSTOR.

Enhance your scholarly research with underground newspapers, magazines, and journals.

Explore collections in the arts, sciences, and literature from the world’s leading museums, archives, and scholars.

• Find journals
• My journals

Register Sign in

Register or sign-in in order to manage your journal lists

Sign in or register to save a journal

To save a journal and create lists, you need to sign in to your Elsevier account.

Find the right journal for your research

Looking for the best journal match for your paper? Search the world's leading source of academic journals using your abstract or your keywords and other details.

Check if you're eligible for open access (OA) savings.

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here .

Loading metrics

Open Access

Peer-reviewed

Research Article

How do soundboard-trained dogs respond to human button presses? An investigation into word comprehension

Roles Conceptualization, Data curation, Investigation, Methodology, Project administration, Writing – original draft, Writing – review & editing

Affiliations Department of Cognitive Science, University of California, San Diego, San Diego, California, United States of America, Department of Psychological & Brain Sciences, Johns Hopkins University, Baltimore, MD, United States of America

Roles Conceptualization, Investigation, Methodology

Affiliations College of Arts and Sciences, Canisius College, Buffalo, New York, United States of America, FluentPet, Inc, San Diego, California, United States of America

Roles Investigation, Methodology, Project administration, Resources, Software

Affiliations Department of Cognitive Science, University of California, San Diego, San Diego, California, United States of America, Department of Psychological & Brain Sciences, Johns Hopkins University, Baltimore, MD, United States of America, FluentPet, Inc, San Diego, California, United States of America

Roles Formal analysis, Visualization, Writing – review & editing

Affiliation Department of Linguistics, University of California, Davis, Davis, California, United States of America

Roles Formal analysis, Writing – review & editing

Affiliations FluentPet, Inc, San Diego, California, United States of America, Statistics and Operational Research Department, Universitat de València, Valencia, Spain

Roles Conceptualization, Methodology, Writing – review & editing

Affiliations Comparative Cognition, Messerli Research Institute, University of Veterinary Medicine Vienna, Medical University of Vienna, University of Vienna, Vienna, Austria, School of Psychology and Neuroscience, University of St. Andrews, St. Andrews, United Kingdom

Roles Investigation

Affiliation Department of Cognitive Science, University of California, San Diego, San Diego, California, United States of America

Affiliations Department of Cognitive Science, University of California, San Diego, San Diego, California, United States of America, FluentPet, Inc, San Diego, California, United States of America

Roles Conceptualization, Writing – review & editing

Roles Conceptualization, Methodology, Project administration, Resources, Supervision, Writing – review & editing

* E-mail: [email protected]

• Amalia P. M. Bastos, 
• Ashley Evenson, 
• Patrick M. Wood, 
• Zachary N. Houghton, 
• Lucas Naranjo, 
• Gabriella E. Smith, 
• Alexandria Cairo-Evans, 
• Lisa Korpos, 
• Jack Terwilliger, 

• Published: August 28, 2024
• https://doi.org/10.1371/journal.pone.0307189
• Peer Review
• Reader Comments

Past research on interspecies communication has shown that animals can be trained to use Augmentative Interspecies Communication (AIC) devices, such as soundboards, to make simple requests of their caretakers. The recent uptake in AIC devices by hundreds of pet owners around the world offers a novel opportunity to investigate whether AIC is possible with owner-trained family dogs. To answer this question, we carried out two studies to test pet dogs’ ability to recognise and respond appropriately to food-related, play-related, and outside-related words on their soundboards. One study was conducted by researchers, and the other by citizen scientists who followed the same procedure. Further, we investigated whether these behaviours depended on the identity of the person presenting the word (unfamiliar person or dog’s owner) and the mode of its presentation (spoken or produced by a pressed button). We find that dogs produced contextually appropriate behaviours for both play-related and outside-related words regardless of the identity of the person producing them and the mode in which they were produced. Therefore, pet dogs can be successfully taught by their owners to associate words recorded onto soundboard buttons to their outcomes in the real world, and they respond appropriately to these words even when they are presented in the absence of any other cues, such as the owner’s body language.

Citation: Bastos APM, Evenson A, Wood PM, Houghton ZN, Naranjo L, Smith GE, et al. (2024) How do soundboard-trained dogs respond to human button presses? An investigation into word comprehension. PLoS ONE 19(8): e0307189. https://doi.org/10.1371/journal.pone.0307189

Editor: Brenton G. Cooper, Texas Christian University, UNITED STATES OF AMERICA

Received: April 18, 2024; Accepted: July 1, 2024; Published: August 28, 2024

Copyright: © 2024 Bastos et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: Study pre-registration is deposited in the Open Science Framework, https://osf.io/tskcq Data and analysis scripts are available from Github, https://github.com/znhoughton/comprehension .

Funding: A.P.M.B. is supported by the Johns Hopkins Provost’s Postdoctoral Fellowship Program.

Competing interests: A.P.M.B., P.M.W., Z.N.H., J. T., G.E.S., & L.K. have previously consulted for FluentPet, Inc., a company that produces AIC devices for pets. A.E. & L.N. are employees of FluentPet, Inc.

Introduction

The use of Augmentative Interspecies Communication (AIC) devices has gained popularity among pet owners in recent years, especially among dog owners [ 1 – 3 ]. Most of these AIC devices consist of soundboards with buttons that, when pressed, produce pre-recorded words in the owner’s voice. Most owners report that they train their dogs using a technique that has become known as modelling: owners demonstrate the actions associated with each button repeatedly until their animals make presses themselves, at which point the same actions are performed by the owners, regardless of whether the animal’s press was intentional or accidental [ 2 ]. Over time, animals may press the buttons on their soundboards more frequently, which may in turn provide them with more control over their daily lives and environments.

Although interspecies communication through soundboard-like AIC devices has been adopted to varying degrees of success with apes, dolphins, and other species (for a review, [ 1 ]), this has primarily been undertaken by researchers and professional animal trainers. In the scientific literature, only a single dog named Sofia has been shown to use buttons to request actions such as going on a walk or playing, following training by a professional dog trainer [ 4 , 5 ]. The training methods used with Sofia are considerably different from the modelling approach that is most common among pet owners: Sofia was first taught to tap her AIC device with her paw on command, which was eventually phased out once she began doing so spontaneously [ 4 ]. Given the nature of Sofia’s training, it is also possible that Sofia experienced less cueing from her trainer than other owner-trained pet dogs, whose performance may be more susceptible to the Clever Hans effect [ 6 , 7 ].

Crucially, communicative production of words, such as that demonstrated by Sofia, must be preceded by the comprehension of said words. Although Sofia is the only dog ever shown to produce requests through an AIC device, many more dogs are capable of responding appropriately to spoken human signals. Most pet dogs are trained to respond to at least a few vocal signals from their owners such as “sit” and “lie down” [ 8 , 9 ], but some dogs such as Rico and Chaser can learn tens or hundreds of names for individual objects ([ 10 – 17 ]; although learning object names may be more challenging to dogs than learning words associated with actions, see [ 18 ]). Given that owner-trained soundboard-using dogs have not yet been tested in controlled experimental contexts, we currently do not have any evidence to suggest that they have successfully associated the words produced by their soundboards to their respective consequences, let alone that they produce said words communicatively.

Another potential issue surrounding owner-trained soundboard-using dogs concerns the reliability of citizen science data [ 19 ]. Although thousands of pet owners currently contribute regular data on their animals’ soundboard use to scientific research [ 2 ], the extent to which reported interactions are cued, inaccurately reported, or cherry picked by owners is unclear and a matter under current investigation. Consequently, testing soundboard-trained dogs under controlled experimental conditions is a necessary step towards assessing this form of interspecies communication.

Given the novelty of this approach, the nature of most of the data on this form of interspecies communication, and our limited knowledge of dogs’ capacity to acquire associations in this context, in the present study we pursue a thorough and multifaceted investigation into the comprehension of button presses by humans. We begin from the premise that owner-trained soundboard-using dogs are probably more likely to associatively learn the words for routine activities or events (such as learning that the word “outside”, when spoken or pressed by a human, is usually followed by a door being opened to the backyard) than some of the more abstract words sometimes provided on owner-trained dogs’ soundboards (such as “tomorrow” or “later”). Given that dogs’ soundboards are determined by their owners and therefore are composed of buttons with different words, and have different layouts (for a detailed explanation, see [ 2 ]), here we select three of the most commonly occurring routine-related word concepts across all participants (namely, food-related, play-related, and outside-related words) and choose the most common variants within these concepts (OUT or OUTSIDE for outside-related words, PLAY or TOY for play-related words, and FOOD or EAT or DINNER or HUNGRY for food-related words) for our experiments (as pre-registered in the Open Science Framework: https://osf.io/tskcq ). Over the course of two complementary experiments, we use these three word concepts, as well as a nonsense word (the nonce word “DAXING”) to answer five fundamental questions about owner-trained soundboard-using dogs’ comprehension of the words on their soundboards.

First, using both a researcher-led in-person experiment and a remotely conducted citizen science experiment, we test whether dogs can recognise and respond appropriately to the words recorded onto their soundboards. We hypothesise that, if dogs have associatively learnt the connection between these words and the outcomes they usually entail, then they should behave in anticipation of the actions or events indicated by these words. For example, upon hearing the word “PLAY”, dogs should be more likely to walk to their toy box and pick up a toy, than if they had heard “FOOD”. Conversely, we should see fewer looks to a food bowl following the word “PLAY” than following “FOOD”. Given that these words might typically be produced by owners concurrently with many other contextual cues (for example, “FOOD” might be produced most commonly at specific mealtimes, while the owner fills up the dog’s bowl), in our experimental trials we strip all words of any additional contextual cues, to ensure that dogs’ responses occur specifically in response to the words themselves.

Second, we use the two experiments to compare dogs’ responses to a word when it is produced by the owner and by an unfamiliar person (a researcher). We hypothesise that, if dogs expect words to precede their accompanying actions, and can do so without any additional contextual or owner-provided cues, then they should respond similarly regardless of the person producing the word. Given that words on dogs’ buttons usually consist of recordings of their owner’s speech (as was the case for every dog participating in this study), and that differences in the acoustic properties of different people’s speech can conceivably generate differences in dogs’ behavioural responses unrelated to their comprehension of words, we compare their responses to button presses by the owner and presses of those same buttons by an unfamiliar person.

Third, a comparison between the two experiments–particularly in terms of dogs’ responses to button presses by their owners and an unfamiliar person–can also inform whether the results of citizen science studies on soundboard-trained dogs are broadly comparable to in-person researcher-led studies, and therefore establish whether future work with this population could be carried out through data collected remotely by citizen scientists.

Fourth, we also compare dogs’ responses to owners’ button presses and owners’ vocal production of these same words. If dogs respond appropriately and equivalently to words regardless of their mode of production (whether they are spoken directly by the owner or produced by a button press), then this would suggest that dogs are associating words’ consequences (e.g., the actions following words, such as a play interaction following “PLAY”) to the words (speech or speech recordings) themselves, rather than attending primarily to other cues such as the location of the buttons (which would predict contextually-appropriate responses when buttons are pressed, but not when the same words are spoken out loud). On the other hand, if dogs respond appropriately to owners’ button presses of words but not to their vocal production, then it is likely that dogs associate button locations–or some other property of the buttons–to their accompanying actions or events, disregarding button audio.

Finally, because of recent research suggesting that dogs tilt their heads in response to familiar words more so than unfamiliar ones [ 20 ], and that this is a lateralized behaviour, we investigated subjects’ head tilts throughout the two experiments. We hypothesise that, if dogs do in fact tilt their heads in response to words that they recognise, and soundboard-trained dogs recognise the words recorded onto their buttons, then we should see more head-tilting behaviour in response to the known words (i.e., outside-related, play-related, and food-related words) than to a nonce word (“DAXING”). This result should hold regardless of whether words are produced by button presses or spoken by their owner. If this is true, then this study can also provide a larger sample size to inform whether this head-tilting behaviour is more commonly left-lateralized or right-lateralized, or, as in the original study, this is split evenly across subjects. If, on the other hand, we find evidence to suggest that soundboard-trained dogs do respond appropriately to known words but do not tilt their heads upon hearing them, this might instead suggest that there are population or breed differences between soundboard-trained dogs and the “gifted world learner” [ 15 ] dogs in the study conducted by Sommese and colleagues [ 20 ].

In sum, the present study aims to investigate five principal questions: (1) whether dogs can recognise and respond appropriately to the words recorded on their soundboards; (2) whether they exhibit these responses even in the absence of other contextual or owner-produced cues, even when they are produced by an unfamiliar person; (3) whether citizen science studies and in-person studies can produce comparable results in this population of subjects; (4) whether dogs attend to speech specifically in this context, or other cues; and (5) whether owner-trained soundboard-using dogs tilt their heads in response to familiar words in the same way that dogs trained to recognise object names do.

Experiment 1: In person study

For the in-person study (IPS), subjects were 30 dogs (14 males, age: M = 3.52 years, SD = 2.91; see Table 1 for subject information). No subjects were excluded from analyses. All dogs were family pets trained by their owners to use soundboard AIC devices, and whose soundboards contained the words OUT/OUTSIDE, PLAY/TOY, and FOOD/EAT/DINNER/HUNGRY. Owners were asked to habituate their dogs to people wearing sunglasses, hats, and other head accessories prior to the study date, in their own time, given that the study would involve owners wearing multiple accessories on their faces. Dogs were tested in their home settings with their owners present; owners were not told the purpose of the study until testing was concluded. All subjects lived primarily indoors and were fed indoors and the study received ethics approval from the UCSD IACUC (protocol no. S21098).

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

Age at time of testing is provided in years, rounded to the nearest decimal point, as per owner reports. For mixed breed dogs, up to three primary component breeds (based on owner reports or genetic testing) are provided in brackets.

https://doi.org/10.1371/journal.pone.0307189.t001

The study was conducted in dogs’ homes, in the room where their soundboard was typically located. Upon arrival, one researcher (Experimenter 1, hereafter E1) waited outside the house whilst the other (Experimenter 2, hereafter E2) went into the home and greeted the owner and their dog, before placing the dog either in the backyard or in a room in the house other than the one where the soundboard was located.

Once the dog was out of sight, E2 placed three large coloured stickers over the three buttons containing one of each of the words of interest: (1) “OUT” or “OUTSIDE”; (2) “PLAY” or “TOY”; and (3) “FOOD” or “EAT” or “DINNER” or “HUNGRY”. This ensured that, at test, E1 could not tell the buttons’ identity based on any writing or symbols visible on the buttons. E2 then asked that the owner record their voice onto a new button (that matched the size and style of the dog’s existing soundboard buttons) for the word “DAXING” and placed that on the soundboard, with another coloured sticker. Sticker colours were randomly assigned to buttons between subjects, such that E1 would not have known the identity of any buttons based on the colour of the stickers placed on them. Stickers were made from green, red, yellow, and blue cardstock paper and adhesive putty. Owners were shown a novel action–placing their hands on their head and spinning in a slow circle–which they were told they would have to perform at some point later in the study.

The dog was then brought back into the room where their soundboard was located, and E1 was invited into the home and allowed to greet the dog and owner. E2 then set up the camera recording equipment (Panasonic Full HD Camcorder HC-V180K, GoPro Hero7) in the room, and an Anmeate Baby Monitor facing the soundboard, such that multiple camera angles of the room were recorded ( Fig 1 ).

The door used for the button concept “OUTSIDE” is outlined in blue, the dog’s toys are outlined in orange, and the food bowl is outlined in green. During trials for this subject, for example, the owner sat in the reclining chair on the bottom left of the frame, facing away from the soundboard, while wearing noise-cancelling headphones, a face mask (as a Covid-19 precaution), and a black sleep mask.

https://doi.org/10.1371/journal.pone.0307189.g001

While E1 explained the experimental procedure to the owner and ensured the dog was comfortable and behaving normally in their presence, E2 identified a different room in the house to hide so they could remotely monitor the study (through a live feed of the baby monitor). Next, the owner was asked to wear a sleep mask and noise-cancelling headphones and ensure that the dog was not fearful of either accessory. Finally, E2 handed E1 their headphones and set both headphone volumes such that that neither E1 nor the dog’s owner could hear E2 speaking loudly from a short distance, therefore ensuring that they would also be unable to hear the recordings on any of the buttons when they were pressed during a trial.

Trials began with E1 and the owner sitting in the soundboard room whilst the dog ranged freely around the room. The owner wore a sleep mask and listened to music on noise-cancelling headphones so they could not see any of the trial procedures. This ensured that owners were blind to the study’s hypotheses and predictions but were still present in the room to ensure that dogs were comfortable and behaving normally.

At the start of each trial, E2 remotely triggered a sound file to E1’s headphones stating the colour of the sticker for the button they would press in that trial, then remotely started playing music into the headphones of both E1 and the dog’s owner. E1 looked straight ahead until they heard a beep in their headphones, again triggered via Bluetooth by E2. At this point, E1 stood up and began the trial: E1 walked to the same place by the soundboard each time, pressed the button covered by the sticker colour they were given, stepped away from the button again, and looked in the direction of the dog. E1 froze in this position for 1 minute while they waited for the next beep in their headphones, at which point they turned around to face the nearest wall and closed their eyes. This final beep and turn concluded the trial.

Following the end of the trial, E2 triggered an audio file in the owner’s headphones stating that they could remove their headphones and sleep mask and gave the owner instructions on an action to perform. Unbeknownst to the owner, this action always matched the button recently pressed by E1. For “OUT” or “OUTSIDE”, the owner was asked to let their dog outside as they normally would (e.g., open the door to the backyard so the dog could go out). For “PLAY” or “TOY”, the owner was asked to briefly engage their dog in play with a toy. For “FOOD”, “EAT”, “DINNER”, or “HUNGRY”, the owner was asked to place a small amount of their dog’s usual food in their food bowl. For “DAXING”, the owner performed the novel action they practiced earlier, namely placing both hands on the top of their head and spinning in a slow circle. While the owner performed their action, E1 remained with their eyes closed, listening to music through noise-cancelling headphones, so that they would not be unblinded to the button they just pressed. Once the owner performed their action, E2 came out of their hiding place and asked the owner to return the room to its starting state (e.g., closing any doors that were opened, placing any toys back where they were at the start of the trial), and returning to their starting position in the room. E2 then walked over to E1 and tapped them on the shoulder, so that E1 knew they could remove their headphones, open their eyes, and turn back around.

Experiment 2: Citizen science study

In the citizen science study (CSS), subjects were a new group of 29 dogs (9 males, age: M = 2.95 years, SD = 1.91; see Table 2 for subject information). Another 3 subjects (3 males) completed the study but had their data excluded due to procedural errors, and a further 8 subjects (4 males) were recruited for the study but never completed it. All dogs were owner-trained family pets with soundboards containing the same words as required in Experiment 1. Owners were asked to habituate their dogs to sunglasses prior to beginning the study, since owners were asked to wear mirrored sunglasses throughout all trials. Dogs were tested in their home settings by their owners, who were not told the purpose of the study until testing was concluded. As before, all dogs lived primarily indoors and were fed their meals indoors, and the study received ethics approval from the UCSD IACUC (protocol no. S21098).

https://doi.org/10.1371/journal.pone.0307189.t002

Owners were given study instructions in three formats: an explanatory video demonstrating study procedures; written materials with step-by-step instructions; and a 15-minute zoom call with a researcher who talked them through the study instructions and answered any remaining questions. As with the in-person study, owners were not told the purpose of the experiment until after they concluded their participation and submitted video data of all trials to the research team. Debriefs were also done by an experimenter through another 15-minute zoom call. Given that they administered the experimental trials to their own dogs, owners could not be made blind to conditions.

In each trial, owners either pressed the button for, or spoke, one of the words used in Experiment 1, namely: (1) “OUT” or “OUTSIDE”; (2) “PLAY” or “TOY”; and (3) “FOOD” or “EAT” or “DINNER” or “HUNGRY”, and (4) the nonce word “DAXING”. Each word was produced twice by the owner over the course of the study, once spoken and once pressed. Owners carried out two trials per day, allowing an interval of at least 30 minutes between the two trials within a day, over the course of four days, which were spread over a maximum of 2 weeks. Therefore, dogs in this study experienced a total of eight conditions: a spoken outside word, a pressed outside word, a spoken play word, a pressed play word, a spoken food word, a pressed food word, the word “DAXING” produced by a button press, and “DAXING” spoken by their owner. Trial orders were randomised across subjects.

Before the start of each trial, owners were asked to move their dog to a different room of the house where they could not see their soundboard. Owners then recorded a new button for the word “DAXING” and placed it in an empty slot on their dog’s soundboard. The dog was brought back into the room where their soundboard was located for the start of the trial.

Each trial began with the owner standing in the same location next to the dog’s soundboard. They then either pressed the button or spoke the word specified for that trial. Owners then refrained from interacting with their dog for the next minute; owners were specifically instructed not to speak to, point to, react to, gesture at, or otherwise interact with their dog or their environment for a full minute. At the end of each trial, owners were asked to perform the matching action for the pressed or spoken word, as specified by their participant sheet (i.e., as in Experiment 1, owners were asked to open the door and let their dog into the backyard following the production of the word “OUTSIDE”, and spin slowly with their hands on their heads following the word “DAXING”). Following their post-trial action, owners again took their dogs out of the room and returned the soundboard to its usual configuration, removing the button for the word “DAXING”, such that dogs could not press the “DAXING” button in the intervals between trials and on non-trial days. The full trial time was recorded on video by the owners, sometimes from multiple camera angles.

Coding and analyses.

Videos for both studies (Experiment 1 and Experiment 2) were processed prior to blind coding to ensure that blind coders could never observe the button press produced by E1 (Experiment 1) or the button press or word spoken by the owner (Experiment 2), as well as the post-trial action performed by the owner. Video files were renamed so as to remove any trial information. Videos from different camera angles of the same trial were time-matched and aligned side-by-side into a single video file. E1 and the owner were occluded by a black rectangle such that blind coders could not see any humans on the screen, and the video was cut so that it began following the moment that E1 stood upright after pressing a button and ended 60 seconds later.

Blind coders were naïve to experimental design, experimental hypotheses, and conditions. They were initially trained on a set of videos which included trials of the study piloted with non-eligible dogs and non-study videos crowdsourced online of non-subjects. Coders annotated videos following a pre-determined ethogram ( Table 3 ). Interrater agreement was substantial throughout training (Kappa coefficient of 0.81, with a confidence interval of [0.76; 0.86] for 100% of training set) and remained satisfactory throughout the coding of the full dataset (Kappa coefficient of 0.70, with a confidence interval of [0.64; 0.76] for a randomly selected subset comprising 10% of all data).

We predicted that dogs would show more outside-directed behaviours in response to “OUT” or “OUTSIDE” words, more toy-directed behaviours in response to “PLAY” or “TOY” words, and more food-directed behaviours in response to “FOOD”, “EAT”, “DINNER”, or “HUNGRY” words.

https://doi.org/10.1371/journal.pone.0307189.t003

To determine whether dogs responded as expected to E1’s button presses (in Experiment 1) and the owners’ button presses (in Experiment 2), we investigated whether their behavioural responses to button presses were consistent with individual buttons’ words. For example, upon observing a button press for “OUTSIDE”, a dog should be more likely to move toward the door than upon observing a press for “FOOD”.

The data was analysed using Bayesian linear mixed-effects models implemented in brms [ 21 ] in R [ 22 ]. Specifically, we used a Bernoulli model which requires the dependent variable to be binary. In our case, we ran three separate Bernoulli models, one for each type of behaviour (food-directed behaviours, outside-directed behaviours, and play-directed behaviours). Our reasoning was that if dogs have created associations between these words and their meanings, then they should show the appropriate behaviour associated with that button in the condition.

The dependent variable for each model was whether, on any given trial, the dog showed the target behaviour. For example, in the model for the FOOD condition, the dependent variable was 1 for a given trial if the dog showed a food-related behaviour, and 0 if not (where “trial” here refers to a coded behaviour, so each food-related behaviour would be coded 1 in the FOOD model, and each non-food behaviour would be coded 0). A complete model description can be found in the analysis script (which is provided in full at https://github.com/znhoughton/comprehension ).

The independent variables were the three conditions (Food Condition, Outside Condition, or Play Condition), Experiment (IPS or CSS), and the interaction between the two, with maximal random effects [ 23 ]. For each model, we used weakly informative priors. The model syntax for each model is included below:

• Model 1: Play behaviours ~ Condition*Experiment + (1 + Condition|Subject)
• Model 2: Outside behaviours ~ Condition*Experiment + (1 + Condition|Subject)
• Model 3: Food behaviours ~ Condition*Experiment + (1 + Condition|Subject)

Next, using only data from Experiment 2 (CSS), we tested whether dogs behave comparably in response to a spoken word and a word produced by a button press. To do so, we once again used three Bayesian logistic regression models implemented in brms in R, the same dependent variable as above, but with the fixed-effects of condition and mode (i.e., spoken or pressed; and their interaction), random slopes for condition by subject. This is described in the following three equations:

• Model 4: Food behaviours ~ Condition*Mode + (1 + Condition*Mode|Subject)
• Model 5: Outside behaviours ~ Condition*Mode + (1 + Condition*Mode|Subject)
• Model 6: Play behaviours ~ Condition*Mode + (1 + Condition*Mode|Subject)

Finally, given recent work to suggest head-tilting is a lateralized behaviour which occurs when dogs recognise and process known words [ 18 ], we planned to compare the number of head tilts (to both sides first, then separately for the right and left sides) dogs made in the nonce (“Daxing”) condition compared to the three meaningful known-word conditions (Outside, Play, and Food conditions) in both the researcher-led experiment and the citizen science component of this study. In order to assess this, we set out to use a Bayesian linear mixed-effects model with a Poisson distribution implemented in brms in R, with number of head tilts as the dependent variable and condition and context (known word vs. nonce word) as well as their interaction as fixed-effects and random intercepts for subject and random slopes for condition by subject. Our model specification is listed in the following equation: Number of head tilts ~ Condition * Experiment + (1|Subject) + (Condition|Subject).

All analyses were pre-registered in the Open Science Framework ( https://osf.io/tskcq ).

Question 1: Can dogs recognise and respond appropriate to the words recorded on their soundboards?

Across both experiments, dogs exhibited approximately seven times more play-directed behaviours in the Play Condition ( Table 4 ), and approximately seven times more outside-directed behaviours in the Outside Condition ( Table 5 ), suggesting that they recognised and responded appropriately to these two words. We found no conclusive evidence to suggest that dogs exhibited food-directed behaviours in the Food Condition compared to the other two conditions ( Table 6 ). Note that, since all models used sum coding, the intercept in all tables represents the grand mean, and the coefficient values represent the distance, in log-odds, between the effect and the intercept. Dogs’ behaviours across the three familiar-word conditions of both experiments are shown in Fig 2 .

Proportion here is the number of behaviours of a specific type (e.g., outside behaviours) in a given condition (e.g., play condition) divided by the total number of behaviours in that condition. The x-axis is Condition. The y-axis is the average proportion of each behaviour type across participants. Each of the points represents the proportion of behaviours of the corresponding type of behaviour in the respective condition. The error bars represent standard errors.

https://doi.org/10.1371/journal.pone.0307189.g002

In the Play Condition, the odds of dogs displaying play-directed behaviours was approximately seven times greater than average across the three meaningful-word conditions.

https://doi.org/10.1371/journal.pone.0307189.t004

In the Outside Condition, the odds of dogs displaying outside-directed behaviours was approximately seven times greater than average across the three meaningful-word conditions.

https://doi.org/10.1371/journal.pone.0307189.t005

Dogs were not more likely to perform food-directed behaviours in the Food Condition compared to Play or Outside Conditions.

https://doi.org/10.1371/journal.pone.0307189.t006

Question 2: Do dogs respond equivalently to words produced by their owner compared to an unfamiliar person?

We found no effect of button presser identity (owner or unfamiliar person) on dog’s behaviours for any of the three conditions (see confidence intervals overlapping zero in Tables 5 – 7 for IPS and CSS comparisons). This suggests that dogs respond appropriately to button presses even in the absence of other contextual cues or owner-produced cues.

“Play:Press”, “Food:Press” and “Outside:Press” correspond to the interactions between each condition and mode. For example, the effect of “Play:Press” is the extent to which the effect of “Play” differs when the dog hears a recording from a button press as opposed to a spoken word.

https://doi.org/10.1371/journal.pone.0307189.t007

Question 3: Can citizen science studies and in-person studies of soundboard-trained dogs produce comparable results?

We found no difference in dogs’ behaviours in response to owner-produced and experimenter-produced button presses (see Tables 4 – 6 ), and therefore the results for both experiments were comparable. In the present investigation, dogs behaved similarly in response to button presses in the in-person study (IPS) and the citizen science study (CSS), suggesting that both experiments produced equivalent results.

Question 4: Do soundboard-trained dogs attend to speech?

Within Experiment 2 (IPS), dogs’ responses to spoken and pressed words across the three conditions were comparable for food-related behaviours ( Table 7 ), outside-related behaviours ( Table 8 ), and play-related behaviours ( Table 9 ). Overall, these results suggest that the two modes of word-production led to equivalent behavioural responses by subjects.

https://doi.org/10.1371/journal.pone.0307189.t008

https://doi.org/10.1371/journal.pone.0307189.t009

Question 5: Do soundboard-trained dogs tilt their heads in response to familiar words? Ultimately, our dataset contained too few head-tilting observations (9 instances) for this analysis to be completed, so we cannot ascertain whether head-tilting was more common in meaningful known-word conditions (Play, Food, and Outside Conditions) compared to the nonce word condition (Daxing Condition).

Our study suggests that dogs were more likely to perform play-related behaviours after an experimenter or their owner produced a play-related word, and were more likely to exhibit outside-related behaviours in response to an experimenter or their owner producing an outside-related word. This demonstrates that dogs are, at the very least, capable of learning an association between these words or buttons and their outcomes in the world. On the other hand, dogs did not produce preferentially food-related behaviours in response to the relevant words. It is possible that this may have occurred because several of the dogs taking part in the study may have been satiated before the start of test trials, or because dogs did not expect that they would be served a meal outside of their usual feeding times. Our in-person study was typically conducted during work hours, which do not typically overlap with adult dogs’ mealtimes in the early mornings and/or evenings.

Given that dogs responded equivalently to their owners’ button presses and an unfamiliar person’s button presses, our results also demonstrate that dogs attend to and respond to the buttons or words themselves, rather than behaving solely based on unrelated unintentional cues provided by their owners (e.g., as would be expected from a Clever Hans effect). In both experiments, even when all word-related contextual cues were removed from the human’s interaction with the AIC device, dogs still responded with contextually appropriate behaviours. For example, if dogs typically observe their owners picking up a toy before they press a button for a play-related word, any play-related behaviours the dogs exhibit in turn might occur in response to the sight of the toy, the owner’s button press, or a combination of both events. Since our citizen science experiment specifically required that owners press buttons without performing any other actions, the fact that dogs still displayed contextually appropriate responses in the absence of other cues demonstrates that they attended specifically to the words recorded onto their buttons. Additionally, dogs responded appropriately to an unfamiliar human making button presses, even though that person was unaware of which button they were pressing.

Both word sounds, and the location of their respective buttons on dogs’ soundboards, are highly correlated with their effects in the world. For example, when owners press the button OUTSIDE, dogs might equally attend to the location of that button on their AIC device, or the sound (word) produced by the button press. Our citizen science study addressed this question by comparing dogs’ responses to their owners’ speech and button presses. We found no differences in dogs’ responses to either mode of word production, suggesting that most dogs do not associate button location alone with the outcomes of button presses. If they did, they should perform considerably more context-appropriate behaviours in response to the relevant button presses than to owners’ spoken words. We note that, although this does not necessarily preclude that dogs may have formed two separate associations–one for the button location and another for the speech sound of the word–, it nevertheless demonstrates that most dogs can and do attend to and appropriately respond to the auditory properties of words.

Crucially, we found no differences in dogs’ behaviours across the two experiments, suggesting that owners’ and researchers’ conduct of the methods was sufficiently equivalent to yield comparable results. This suggests that results from the citizen science version of the study were comparable to those performed by researchers during visits to owners’ homes. This is important because soundboard-trained dogs are spread all over the world and we do not yet know how moving their soundboards outside of their home environment might affect their soundboard use [ 2 ]. Our findings offer a promising outlook for future citizen science studies with this population of owners and their dogs: remotely conducted citizen science studies could be a critical tool for studying this large and geographically widespread population and maintaining long-term engagement of pet owners with this research program. However, we advise that other future studies should again validate citizen science methods by comparing them against researcher-led in-person experiments, therefore replicating our findings, before scientists can rely more heavily on citizen science for this and similar studies. Additionally, we highlight that all procedures and analyses of this study were pre-registered in advance of data collection, and that future studies extending on this work would also benefit from pre-registration. Pre-registration offers an important tool for ensuring transparency and reproducibility of research, and its use is critical to large scale exploratory studies and citizen science studies.

In sum, our findings provide the first evidence of button word comprehension by owner-trained soundboard-using dogs, and demonstrate that dogs’ contextually appropriate responses to button presses were comparable regardless of the identity of the person using the soundboard, and the absence of other environmental cues related to that word. Our findings also suggest that dogs attend to the sounds recorded onto their buttons, given that they responded equivalently to words when they were produced by button presses and when they were spoken by their owners. A summary of our findings is tabulated in Table 10 below.

https://doi.org/10.1371/journal.pone.0307189.t010

In our study, dogs responded to spoken or pressed food-related and outside-related words with contextually appropriate responses slightly more often than expected by chance. The accuracy of dog’s responses in our study is comparable to dogs’ accuracy in responding to human pointing [ 24 ]. However, dogs are capable of much greater accuracy when purposely trained to respond to stimuli, as is the case for dogs trained to detect wildlife [ 25 ], individual people [ 26 ], and chemical substances [ 27 , 28 ] through scent. This discrepancy in performance could be due to the nature of the stimuli: perhaps dogs find it more difficult to form associations between words and their respective outcomes compared to scents and a consistent reward. Alternately, communicative contexts may produce less predictable responses due to the weaker correlation between the perception of relevant stimuli (a pointed finger, or a word) and their outcomes, therefore leading to less strongly conditioned responses to the stimuli. While scent detection dogs undergoing training will likely experience reinforcement almost every single time the target stimulus is present, a pet dog may not receive any reinforcement on a great number of occasions when they hear familiar words or observe human pointing. Relatedly, scent detection dogs’ accuracy is much higher in contexts where target odours are present at very high rates, with implications for their performance in real-world scenarios [ 29 ]. Therefore, it is also possible that task-oriented trained contexts, such as scent detection, are much more motivating to dogs, and therefore more likely to trigger consistent responses, than day-to-day contexts involving more variable reinforcement. Further, measures of performance in working dogs such as scent detection dogs are typically based on a small subset of the population that passes stringent standards of training [ 30 ], and typically involves animals that are selectively bred for the purpose of scent detection and further selected based on temperament or cognitive traits [ 30 – 32 ], whereas the soundboard-trained population comprises owner-trained pet dogs, whose temperaments and cognitive traits are likely to vary considerably.

Having established that soundboard-trained dogs can and do attend to and comprehend words, future work should also disambiguate the extent to which spatial information about button positions, or other potential cues for button identity, might aid dogs’ ability to use AIC devices. Additionally, more research is needed investigating soundboard-trained dogs’ responses to a wider range of words, particularly in comparison to a population of non-soundboard-trained pet dogs. Although owners anecdotally report that owner-trained pet dogs spontaneously acquire comprehension of large spoken vocabularies [ 9 ], there are no fully controlled experiments investigating whether dogs exhibit contextually appropriate spontaneous responses to familiar words in the absence of other contextual cues, as in the present study. This is crucial because, although owners may, for example, report that their dogs respond appropriately to a food-related word, this word is typically presented alongside a myriad of other confounding cues, such as the time of day when the dog’s meals are served, the presence of a bowl, or the behaviours their owner might perform before serving their dog’s dinner. Finally, our ongoing work is investigating dogs’ word production [ 2 ]. In order to determine whether dogs’ performance at word comprehension is reflected in their button pressing, carefully controlled future studies must investigate whether dogs can spontaneously produce contextually appropriate button presses in experimentally induced situations (as in [ 4 ]). Not only would such a study be helpful in understanding the depth of soundboard-trained dogs’ word comprehension, but it would also establish the extent to which AIC devices can be used for two-way interspecies communication, involving both presses made by the owner for their dog, and by the dog for their owner.

Acknowledgments

We are grateful to all our dog participants and their owners for contributing their time and data to our research project.

• View Article
• PubMed/NCBI
• Google Scholar
• 3. Hunger C. How Stella Learned to Talk: The Groundbreaking Story of the World’s First Talking Dog. New York, NY, US: William Morrow; 2021.
• 7. Sebeok TA, Rosenthal R, editors. The Clever Hans phenomenon: Communication with horses, whales, apes, and people. Annals of the New York Academy of Sciences. 1981;364: 309–309.

• Mission and history
• Platform features
• Library Advisory Group
• What’s in JSTOR
• For Librarians
• For Publishers

Open and free content on JSTOR and Artstor

Our partnerships with libraries and publishers help us make content discoverable and freely accessible worldwide

Search open content on JSTOR

Explore our growing collection of Open Access journals

Early Journal Content , articles published prior to the last 95 years in the United States, or prior to the last 143 years if initially published internationally, are freely available to all

Even more content is available when you register to read – millions of articles from nearly 2,000 journals

Thousands of Open Access ebooks are available from top scholarly publishers, including Brill, Cornell University Press, University College of London, and University of California Press – at no cost to libraries or users.

This includes Open Access titles in Spanish:

• Collaboration with El Colegio de México
• Partnership with the Latin American Council of Social Sciences

Images and media

JSTOR hosts a growing number of public collections , including Artstor’s Open Access collections , from museums, archives, libraries, and scholars worldwide.

Research reports

A curated set of more than 34,000 research reports from more than 140 policy institutes selected with faculty, librarian, and expert input.

Resources for librarians

Open content title lists:

• Open Access Journals (xlsx)
• Open Access Books (xlsx)
• JSTOR Early Journal Content (txt)
• Research Reports

Open Access ebook resources for librarians

Library-supported collections

Shared Collections : We have a growing corpus of digital special collections published on JSTOR by our institutional partners.

Reveal Digital : A collaboration with libraries to fund, source, digitize and publish open access primary source collections from under-represented voices.

JSTOR Daily

JSTOR Daily is an online publication that contextualizes current events with scholarship. All of our stories contain links to publicly accessible research on JSTOR. We’re proud to publish articles based in fact and grounded by careful research and to provide free access to that research for all of our readers.

1. Introduction

2. the ssrf and the big data science center, 3. integrating the imaging beamlines, 5. conclusions, supporting information.

research papers \(\def\hfill{\hskip 5em}\def\hfil{\hskip 3em}\def\eqno#1{\hfil {#1}}\)

Accelerating imaging research at large-scale scientific facilities through scientific computing

a Big Data Science Center, Shanghai Synchrotron Radiation Facility, Shanghai Advanced Research Institute, Chinese Academy of Sciences, No. 239 Zhangheng Road, Shanghai 201210, People's Republic of China, and b Shanghai Synchrotron Radiation Facility, Shanghai Advanced Research Institute, Chinese Academy of Sciences, No. 239 Zhangheng Road, Shanghai 201210, People's Republic of China * Correspondence e-mail: [email protected] , [email protected]

To date, computed tomography experiments, carried-out at synchrotron radiation facilities worldwide, pose a tremendous challenge in terms of the breadth and complexity of the experimental datasets produced. Furthermore, near real-time three-dimensional reconstruction capabilities are becoming a crucial requirement in order to perform high-quality and result-informed synchrotron imaging experiments, where a large amount of data is collected and processed within a short time window. To address these challenges, we have developed and deployed a synchrotron computed tomography framework designed to automatically process online the experimental data from the synchrotron imaging beamlines, while leveraging the high-performance computing cluster capabilities to accelerate the real-time feedback to the users on their experimental results. We have, further, integrated it within a modern unified national authentication and data management framework, which we have developed and deployed, spanning the entire data lifecycle of a large-scale scientific facility. In this study, the overall architecture, functional modules and workflow design of our synchrotron computed tomography framework are presented in detail. Moreover, the successful integration of the imaging beamlines at the Shanghai Synchrotron Radiation Facility into our scientific computing framework is also detailed, which, ultimately, resulted in accelerating and fully automating their entire data processing pipelines. In fact, when compared with the original three-dimensional tomography reconstruction approaches, the implementation of our synchrotron computed tomography framework led to an acceleration in the experimental data processing capabilities, while maintaining a high level of integration with all the beamline processing software and systems.

Keywords: scientific computing ; synchrotron ; imaging ; automation ; tomography .

2.1. The SSRF and the SSRF Phase-II Beamline Project

2.2. the big data science center, 2.3. the imaging beamlines, 3.1. imaging data pipelines.

3.2. Integration of the SR-CT reconstruction applications

The reconstruction application functionalities provide comprehensive support for the full-field nanoscale CT and microscale CT scanning data reconstructions, while supporting the selection of the reconstruction algorithms, including the filtered back-projection (FBP) and the algebraic reconstruction technique (ART). They integrate the ordered subset expectation maximization (OSEM) algorithm, and the capability of performing missing angle reconstruction. Furthermore, they provide GPU-accelerated real-time and efficient 3D reconstruction capabilities, and support for image pre-processing capabilities (background subtraction, filtering, smoothing, etc .), for ring artifact removal, for the batch reconstruction of the micro-CT data and for semi-automatic geometric parameter correction.

Meanwhile, the application GUI functionalities include a client–server architecture, designed to operate within a LAN, and facilitating seamless client–server interactions. Furthermore, they support manual annotation and correction driven by sample-specific features inside the nano-CT projection data; pre-reconstruction of the projection data, allowing a quick preview of the reconstruction results; region of interest reconstructions, enabling the users to select the desired reconstruction area; sample spatial orientation correction, allowing the users to adjust the sample spatial orientation as-needed; and real-time 3D multi-planar reformation (MPR) rendering of the reconstructed data.

3.3. Framework architecture

The system driver layer, which includes the GPU pipeline driver, the pipeline message driver and the Linux Application Programming Interface (API) for the tomography pipeline, runs on the BDSC cluster and is responsible for providing a standard API interface for the GPU task submission pipeline and scheduling management, and for sending task-end messages to the Kafka message management cluster, once a task ends.

The core framework service layer runs on the BDSC data processing node, and it is responsible for receiving and processing pipeline task status messages, initializing the imaging beamline data collection and processing programs, parsing and extracting all the input and output metadata, in real-time, from the imaging beamline reconstruction pipelines, as well as for parsing the metadata from the scheduling system. Based on the management rules set for the SSRF-SciCat scientific metadata system, all the extracted metadata are then persistently stored within the SSRF-SciCat metadata repository.

The client application layer runs on the terminal workstations at the imaging beamlines, and it is responsible for providing the processing pipeline APIs to the integrated SR-CT reconstruction applications. Both the client and the reconstruction applications are deployed together. When the reconstruction applications utilize the BDSC HPC resources for the reconstructions, then the pipeline client is launched, accordingly. It receives input and data from the reconstruction applications, parses the specified parameter files, submits the computational tasks to the BDSC and then returns the status of BDSC data processing and the path to the output data. The reconstructed data can, then, be accessed directly from the imaging beamline workstations. The client application layer supports two operating modes: a Windows client mode and a Linux client mode, with a unified account and storage management system provided for both Windows and Linux systems.

3.4. Functional modules

3.5. framework workflow.

3.6. Data transfer and management

When the input data are located on the local workstation storage, the BDSC SR-CT framework automatically detects and synchronizes them with the reconstruction task directory under the current AD account path on the BDSC. Each reconstruction task directory is named by default after the input file name and a timestamp. On the other hand, when the input data are stored on the BDSC network storage, the BDSC SR-CT framework does not synchronize the data. Instead, it generates a parameter file anew, and updates the parameter path with the actual storage location. The reconstructed data, generated by the reconstruction task, will be then stored inside the output directory, which is located inside the corresponding reconstruction task directory. Users can then directly access their data from the output directory at the beamline workstation.

3.7. Metadata

The pipeline execution procedure is divided into six stages, as illustrated in Fig. S2 of the supporting information , where, from left to right, LOGIN reports the user's login status, CHECK reports the framework input file check status, SUBMIT reports the computing job submission status, QUEUE reports the computing job queuing status, RUN reports the computing job execution status, and DONE reports the accomplishment status of the entire pipeline.

3.9. Framework benefits

The BDSC SR-CT framework provides several advantages and improvements to the imaging experiments at the synchrotron facilities:

(i) Application-agnostic. The system is highly decoupled from the SR-CT reconstruction applications and other systems at the imaging beamline. It is not limited to a specific imaging application, and it can easily integrate a large plethora of different CT reconstruction software. Moreover, only very minimal modifications are required in order to create a completely new automated processing pipeline for a totally different imaging application or task.

(ii) Metadata management. With the deployment of the SR-CT framework, the BDSC established a standardized tagging system for the labeling of the synchrotron imaging metadata parameters. It can automatically parse the parameter and value tags, and then submit tasks and synchronize data, accordingly. Other CT reconstruction applications can, thus, output parameters based on the BDSC SR-CT tagging system specifications. Moreover, the BDSC SR-CT framework is also capable of parsing customized parameter and value tags from other systems and different applications, and convert them into a standard metadata structure, ultimately feeding them into the SSRF-SciCat system.

(iii) SR-CT beamline reconstruction software integration. The framework client is developed using the Java programming language, providing a cross-platform user interface, while being able of integrating other software developed using different programming languages. Real-time data processing and analysis is achieved through inter-process communication via the deployment of process pipelines. This allows seamless integration with any software which could be developed in the future using the same communication method.

(iv) High-performance task processing. The BDSC SR-CT framework integrates the BDSC cluster task submission system and the query interfaces on the server-side, using a cross-platform REpresentational State Transfer (REST) API. The BDSC dynamically schedules tasks based on the resource availability, eliminating the need for the clients to be tightly bound to a specific set of resources, thus fostering scalability and the automatic allocation of the processing resources.

4.1. Improvements of the imaging experimental performances

To assess the performance improvements brought by the SR-CT pipeline framework to the imaging experiments, we conducted performance evaluations in a production environment using the fast X-ray imaging beamline (BL16U2) workstations and a series of dynamic CT experimental datasets. Each set of the dynamic CT raw data consists of 250 (2000 × 1007 pixels) projections, with each projection being 3.84 MB in size. The total size of a single set of the dynamic CT raw data is 960 MB. For each dataset, the reconstructed (2000 × 2000 × 1007 voxels) 3D result data are stored in .RAW files, 7.5 GB in size. Then, these .RAW files can be split into 1007 (2000 × 2000 pixels) slices using .TIFF format, with each slice being 7.63 MB in size.

Results from assessment of the acceleration brought by the BDSC centralized SR-CT pipeline to the synchrotron imaging experimental performances, when compared with the suite, using different numbers of CT datasets and samples CT datasets SR-CT pipeline Acceleration factor 1 90.95 min 11.6 min 7.84 8 727 min 12.25 min 59.34 12 1091 min 23.65 min 46.13 Mg–Al alloy 197 min 19 min 10.37 172 min 18.1 min 9.5 4.2. Case studies To showcase the impact of the BDSC SR-CT framework on the experimental studies employing the SR-CT, case studies, from the typical application area of the SR-CT research, are presented employing the BDSC SR-CT framework. 3D imaging of the pores and defects in the magnesium–aluminium alloy. ( ) 3D reconstruction image, ( ) cross-sectional view through the largest pore present in the alloy, ( ) selected region of interest for slicing within the 3D reconstructed volume, and ( ) corresponding longitudinal view of the 3D distribution and structure of the internal pores, are shown. SR-CT ( ) 3D phase-contrast imaging, and ( ) corresponding frontal 2D reconstructed slice, of a fish head ( ), where the different biological tissue structures can be clearly discriminated. The BDSC has architected, developed and deployed, at the SSRF, an SR-CT pipeline software framework capable of effectively harnessing the BDSC scientific computing resources in order to accelerate and augment large-scale, massively parallelized, 3D CT experimental reconstructions at the imaging beamlines. The imaging beamlines at SSRF have been, in fact, fully integrated within the BDSC scientific computing framework, thus enabling the full automation of the real-time data processing and analysis pipelines and feedback, significantly reducing the time necessary for the users to process and analyze their experimental data. Furthermore, the BDSC has also integrated its SR-CT framework into the SSRF unified authentication management system (which has also been developed by the BDSC), thus enhancing the level of the framework data privacy and security. Moreover, the BDSC has architected and deployed the SR-CT metadata infrastructure at the SSRF, where all the inputs and outputs, including data and parameters, are ingested by the SSRF-SciCat system, and aggregated into the JSON universal file format, thus enabling the training of scientific AI through ML approaches. The limitations presented by the PITRE software, which lacked crucial HPC features, including the lack of parallelization capabilities and a codebase relying heavily on CPUs rather than being optimized for harnessing GPU architectural acceleration, proved to be a significant challenge for the BDSC during the design and integration phases of the imaging experimental pipeline into the BDSC HPC infrastructure. To address this limitation, the BDSC has thus designed, developed and deployed a framework capable of wrapping the PITRE software with a software layer, equipped with a GPU-aware scheduler and massive parallelization capabilities. This framework is able to translate non-HPC routines into HPC-optimized reconstructions, effectively allowing scientific software, not optimized for the HPC architecture, to access all the benefits of a modern scientific computing framework. The performance evaluation of the BDSC SR-CT framework demonstrates the extent of the acceleration induced by the SR-CT experimental data processing and analysis capabilities compared with the traditional beamline CT reconstruction approaches, impacting the application areas of the synchrotron imaging methods. The BDSC SR-CT framework architecture is highly decoupled from the beamline data infrastructure and the other systems; this allows it to be generalizable, adaptable, scalable, application agnostic and modularly expandable, thus facilitating its integration into the most diverse SR-CT software frameworks addressing the most heterogeneous scientific cases at other facilities worldwide, while seamlessly and quickly adapting to the new challenges posed by the evolution of the future SR-CT experiments, without requiring a structural change in its architecture. Figure S1: SSRF beamline layout; Table S1: functional modules of the SR-CT framework; Figure S2: the UI of the SR-CT pipeline framework. DOI: https://doi.org/10.1107/S1600577524007239/ju5063sup1.pdf Imaging results. DOI: https://doi.org/10.1107/S1600577524007239/ju5063sup2.mp4 ‡ These authors contributed equally to this paper. Acknowledgements The authors acknowledge support from the Shanghai Synchrotron Radiation Facility, Shanghai Advanced Research Institute, Chinese Academy of Sciences, through the BDSC project. Conflict of interest The authors declare no competing interests. Data availability The authors confirm that the data supporting the findings of this study are available within the article and its supplementary materials. Funding information The following funding is acknowledged: Youth Innovation Promotion Association of the Chinese Academy of Sciences (grant No. 2022290); National Key Research and Development Program of China (grant No. 2021YFA1601000); Institute of High Energy Physics (grant No. National HEP Science Data Center); Young Scientists in Basic Research of the Chinese Academy of Sciences (grant No. YSBR-096); Shanghai Municipal Science and Technology (grant No. Major Project); National Key Research and Development Program Young Scientist Project (grant No. SQ2023YFA1600032). This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence , which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited. Follow J. Synchrotron Rad. Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript. View all journals Explore content About the journal Publish with us Sign up for alerts Top articles Explore the most downloaded* papers from Scientific Reports in 2023. Featuring authors from around the world, these collections highlight valuable research from an international community spanning the natural and clinical sciences. Top articles from previous years can be found on our Collections page . *Data obtained from SN Insights which is based on Digital Science's Dimensions. Quick links Explore articles by subject Guide to authors Editorial policies USC Libraries Research Guides Organizing Your Social Sciences Research Paper Reading Research Effectively Purpose of Guide Design Flaws to Avoid Independent and Dependent Variables Glossary of Research Terms Narrowing a Topic Idea Broadening a Topic Idea Extending the Timeliness of a Topic Idea Academic Writing Style Applying Critical Thinking Choosing a Title Making an Outline Paragraph Development Research Process Video Series Executive Summary The C.A.R.S. Model Background Information The Research Problem/Question Theoretical Framework Citation Tracking Content Alert Services Evaluating Sources Primary Sources Secondary Sources Tiertiary Sources Scholarly vs. Popular Publications Qualitative Methods Quantitative Methods Insiderness Using Non-Textual Elements Limitations of the Study Common Grammar Mistakes Writing Concisely Avoiding Plagiarism Footnotes or Endnotes? Further Readings Generative AI and Writing USC Libraries Tutorials and Other Guides Bibliography Reading a Scholarly Article or Research Paper Identifying a research problem to investigate requires a preliminary search for and critical review of the literature in order to gain an understanding about how scholars have examined a topic. Scholars rarely structure research studies in a way that can be followed like a story; they are complex and detail-intensive and often written in a descriptive and conclusive narrative form. However, in the social and behavioral sciences, journal articles and stand-alone research reports are generally organized in a consistent format that makes it easier to compare and contrast studies and interpret their findings. General Reading Strategies W hen you first read an article or research paper, focus on asking specific questions about each section. This strategy can help with overall comprehension and with understanding how the content relates [or does not relate] to the problem you want to investigate. As you review more and more studies, the process of understanding and critically evaluating the research will become easier because the content of what you review will begin to coalescence around common themes and patterns of analysis. Below are recommendations on how to read each section of a research paper effectively. Note that the sections to read are out of order from how you will find them organized in a journal article or research paper. 1.  Abstract The abstract summarizes the background, methods, results, discussion, and conclusions of a scholarly article or research paper. Use the abstract to filter out sources that may have appeared useful when you began searching for information but, in reality, are not relevant. Questions to consider when reading the abstract are: Is this study related to my question or area of research? What is this study about and why is it being done ? What is the working hypothesis or underlying thesis? What is the primary finding of the study? Are there words or terminology that I can use to either narrow or broaden the parameters of my search for more information? 2.  Introduction If, after reading the abstract, you believe the paper may be useful, focus on examining the research problem and identifying the questions the author is trying to address. This information is usually located within the first few paragraphs of the introduction or in the concluding paragraph. Look for information about how and in what way this relates to what you are investigating. In addition to the research problem, the introduction should provide the main argument and theoretical framework of the study and, in the last paragraphs of the introduction, describe what the author(s) intend to accomplish. Questions to consider when reading the introduction include: What is this study trying to prove or disprove? What is the author(s) trying to test or demonstrate? What do we already know about this topic and what gaps does this study try to fill or contribute a new understanding to the research problem? Why should I care about what is being investigated? Will this study tell me anything new related to the research problem I am investigating? 3.  Literature Review The literature review describes and critically evaluates what is already known about a topic. Read the literature review to obtain a big picture perspective about how the topic has been studied and to begin the process of seeing where your potential study fits within the domain of prior research. Questions to consider when reading the literature review include: W hat other research has been conducted about this topic and what are the main themes that have emerged? What does prior research reveal about what is already known about the topic and what remains to be discovered? What have been the most important past findings about the research problem? How has prior research led the author(s) to conduct this particular study? Is there any prior research that is unique or groundbreaking? Are there any studies I could use as a model for designing and organizing my own study? 4.  Discussion/Conclusion The discussion and conclusion are usually the last two sections of text in a scholarly article or research report. They reveal how the author(s) interpreted the findings of their research and presented recommendations or courses of action based on those findings. Often in the conclusion, the author(s) highlight recommendations for further research that can be used to develop your own study. Questions to consider when reading the discussion and conclusion sections include: What is the overall meaning of the study and why is this important? [i.e., how have the author(s) addressed the " So What? " question]. What do you find to be the most important ways that the findings have been interpreted? What are the weaknesses in their argument? Do you believe conclusions about the significance of the study and its findings are valid? What limitations of the study do the author(s) describe and how might this help formulate my own research? Does the conclusion contain any recommendations for future research? 5.  Methods/Methodology The methods section describes the materials, techniques, and procedures for gathering information used to examine the research problem. If what you have read so far closely supports your understanding of the topic, then move on to examining how the author(s) gathered information during the research process. Questions to consider when reading the methods section include: Did the study use qualitative [based on interviews, observations, content analysis], quantitative [based on statistical analysis], or a mixed-methods approach to examining the research problem? What was the type of information or data used? Could this method of analysis be repeated and can I adopt the same approach? Is enough information available to repeat the study or should new data be found to expand or improve understanding of the research problem? 6.  Results After reading the above sections, you should have a clear understanding of the general findings of the study. Therefore, read the results section to identify how key findings were discussed in relation to the research problem. If any non-textual elements [e.g., graphs, charts, tables, etc.] are confusing, focus on the explanations about them in the text. Questions to consider when reading the results section include: W hat did the author(s) find and how did they find it? Does the author(s) highlight any findings as most significant? Are the results presented in a factual and unbiased way? Does the analysis of results in the discussion section agree with how the results are presented? Is all the data present and did the author(s) adequately address gaps? What conclusions do you formulate from this data and does it match with the author's conclusions? 7.  References The references list the sources used by the author(s) to document what prior research and information was used when conducting the study. After reviewing the article or research paper, use the references to identify additional sources of information on the topic and to examine critically how these sources supported the overall research agenda. Questions to consider when reading the references include: Do the sources cited by the author(s) reflect a diversity of disciplinary viewpoints, i.e., are the sources all from a particular field of study or do the sources reflect multiple areas of study? Are there any unique or interesting sources that could be incorporated into my study? What other authors are respected in this field, i.e., who has multiple works cited or is cited most often by others? What other research should I review to clarify any remaining issues or that I need more information about? NOTE:   A final strategy in reviewing research is to copy and paste the title of the source [journal article, book, research report] into Google Scholar . If it appears, look for a "cited by" reference followed by a hyperlinked number under the record [e.g., Cited by 45]. This number indicates how many times the study has been subsequently cited in other, more recently published works. This strategy, known as citation tracking, can be an effective means of expanding your review of pertinent literature based on a study you have found useful and how scholars have cited it. The same strategies described above can be applied to reading articles you find in the list of cited by references. Reading Tip Specific Reading Strategies Effectively reading scholarly research is an acquired skill that involves attention to detail and an ability to comprehend complex ideas, data, and theoretical concepts in a way that applies logically to the research problem you are investigating. Here are some specific reading strategies to consider. As You are Reading Focus on information that is most relevant to the research problem; skim over the other parts. As noted above, read content out of order! This isn't a novel; you want to start with the spoiler to quickly assess the relevance of the study. Think critically about what you read and seek to build your own arguments; not everything may be entirely valid, examined effectively, or thoroughly investigated. Look up the definitions of unfamiliar words, concepts, or terminology. A good scholarly source is Credo Reference . Taking notes as you read will save time when you go back to examine your sources. Here are some suggestions: Mark or highlight important text as you read [e.g., you can use the highlight text  feature in a PDF document] Take notes in the margins [e.g., Adobe Reader offers pop-up sticky notes]. Highlight important quotations; consider using different highlighting colors to differentiate between quotes and other types of important text. Summarize key points about the study at the end of the paper. To save time, these can be in the form of a concise bulleted list of statements [e.g., intro provides useful historical background; lit review has important sources; good conclusions]. Write down thoughts that come to mind that may help clarify your understanding of the research problem. Here are some examples of questions to ask yourself: Do I understand all of the terminology and key concepts? Do I understand the parts of this study most relevant to my topic? What specific problem does the research address and why is it important? Are there any issues or perspectives the author(s) did not consider? Do I have any reason to question the validity or reliability of this research? How do the findings relate to my research interests and to other works which I have read? Adapted from text originally created by Holly Burt, Behavioral Sciences Librarian, USC Libraries, April 2018. Another Reading Tip When is it Important to Read the Entire Article or Research Paper Laubepin argues, "Very few articles in a field are so important that every word needs to be read carefully." * However, this implies that some studies are worth reading carefully if they directly relate to understanding the research problem. As arduous as it may seem, there are valid reasons for reading a study from beginning to end. Here are some examples: Studies Published Very Recently .  The author(s) of a recent, well written study will provide a survey of the most important or impactful prior research in the literature review section. This can establish an understanding of how scholars in the past addressed the research problem. In addition, the most recently published sources will highlight what is known and what gaps in understanding currently exist about a topic, usually in the form of the need for further research in the conclusion . Surveys of the Research Problem .  Some papers provide a comprehensive analytical overview of the research problem. Reading this type of study can help you understand underlying issues and discover why scholars have chosen to investigate the topic. This is particularly important if the study was published recently because the author(s) should cite all or most of the important prior research on the topic. Note that, if it is a long-standing problem, there may be studies that specifically review the literature to identify gaps that remain. These studies often include the word "review" in their title [e.g., Hügel, Stephan, and Anna R. Davies. "Public Participation, Engagement, and Climate Change Adaptation: A Review of the Research Literature." Wiley Interdisciplinary Reviews: Climate Change 11 (July-August 2020): https://doi.org/10.1002/ wcc.645]. Highly Cited .  If you keep coming across the same citation to a study while you are reviewing the literature, this implies it was foundational in establishing an understanding of the research problem or the study had a significant impact within the literature [either positive or negative]. Carefully reading a highly cited source can help you understand how the topic emerged and how it motivated scholars to further investigate the problem. It also could be a study you need to cite as foundational in your own paper to demonstrate to the reader that you understand the roots of the problem. Historical Overview .  Knowing the historical background of a research problem may not be the focus of your analysis. Nevertheless, carefully reading a study that provides a thorough description and analysis of the history behind an event, issue, or phenomenon can add important context to understanding the topic and what aspect of the problem you may want to examine further. Innovative Methodological Design .  Some studies are significant and should be read in their entirety because the author(s) designed a unique or innovative approach to researching the problem. This may justify reading the entire study because it can motivate you to think creatively about also pursuing an alternative or non-traditional approach to examining your topic of interest. These types of studies are generally easy to identify because they are often cited in others works because of their unique approach to examining the research problem. Cross-disciplinary Approach .  R eviewing studies produced outside of your discipline is an essential component of investigating research problems in the social and behavioral sciences. Consider reading a study that was conducted by author(s) based in a different discipline [e.g., an anthropologist studying political cultures; a study of hiring practices in companies published in a sociology journal]. This approach can generate a new understanding or a unique perspective about the topic . If you are not sure how to search for studies published in a discipline outside of your major or of the course you are taking, contact a librarian for assistance. * Laubepin, Frederique. How to Read (and Understand) a Social Science Journal Article . Inter-University Consortium for Political and Social Research (ISPSR), 2013 Shon, Phillip Chong Ho. How to Read Journal Articles in the Social Sciences: A Very Practical Guide for Students . 2nd edition. Thousand Oaks, CA: Sage, 2015; Lockhart, Tara, and Mary Soliday. "The Critical Place of Reading in Writing Transfer (and Beyond): A Report of Student Experiences." Pedagogy 16 (2016): 23-37; Maguire, Moira, Ann Everitt Reynolds, and Brid Delahunt. "Reading to Be: The Role of Academic Reading in Emergent Academic and Professional Student Identities." Journal of University Teaching and Learning Practice 17 (2020): 5-12. << Previous: 1. Choosing a Research Problem Next: Narrowing a Topic Idea >> Last Updated: Aug 29, 2024 8:43 AM URL: https://libguides.usc.edu/writingguide 209 IMAGES Best scientific journal covers from 2021 Upto 20% Off on International Journal of Advanced Scientific Research (PDF) HOW TO WRITE RESEARCH ARTICLE FOR A JOURNAL: TECHNIQUES AND RULES Scientific Research Journals(Call for Papers) (PDF) How To Write A Scientific Article For A Medical Journal? Anatomy of a Scholarly Article VIDEO [n] Journal meaning (professional publication) with 5 examples Unlocking the Secrets of arxiv Pulling ideas from the brain Writing advanced scientific articles and research papers with A-To-Z guidelines. || Private Batch || Research Tutorial- Hints for reading scientific journal articles Scientific Publications COMMENTS Google Scholar Google Scholar provides a simple way to broadly search for scholarly literature. Search across a wide variety of disciplines and sources: articles, theses, books, abstracts and court opinions. ScienceDirect.com 3.3 million articles on ScienceDirect are open access. Articles published open access are peer-reviewed and made freely available for everyone to read, download and reuse in line with the user license displayed on the article. ScienceDirect is the world's leading source for scientific, technical, and medical research. PubMed PubMed® comprises more than 37 million citations for biomedical literature from MEDLINE, life science journals, and online books. Citations may include links to full text content from PubMed Central and publisher web sites. ... New articles from highly accessed journals Ann Neurol (3) Cell (1) Cell Metab (1) J Clin Endocrinol Metab (5) Methods ... Research articles Research articles. Filter By: Article Type. All. All; Article (208052) Conference Proceeding (56) ... About the journal. About Scientific Reports Contact Journal policies ... Research articles Read the latest Research articles from Nature. ... View all journals; Search; My Account Login; ... Research Article (564) Science in Europe (54) Scientific Correspondence (4028) ... Home PubMed Central ® (PMC) is a free full-text archive of biomedical and life sciences journal literature at the U.S. National Institutes of Health's National Library of Medicine (NIH/NLM) ... Discover a digital archive of scholarly articles, spanning centuries of scientific research. User Guide Learn how to find and read articles of interest to ... Nature First published in 1869, Nature is the world's leading multidisciplinary science journal. Nature publishes the finest peer-reviewed research that drives ground-breaking discovery, and is read by ... Science Science is a leading outlet for scientific news, commentary, and cutting-edge research. Through its print and online incarnations, Science reaches an estimated worldwide readership of more than one million. Science 's authorship is global too, and its articles consistently rank among the world's most cited research. mission & scope. Home Opening Science. PLOS is a nonprofit, Open Access publisher empowering researchers to accelerate progress in science and medicine by leading a transformation in research communication. Every country. Every career stage. Every area of science. Hundreds of thousands of researchers choose PLOS to share and discuss their work. Wiley Online Library One of the largest and most authoritative collections of online journals, books, and research resources, covering life, health, social, and physical sciences. Research science partner journals. Quick Search anywhere. Enter Search Term. Quick Search in Journals. Enter Search Term. Quick Search in Journals. Enter Search Term. ... Science; Research Article or Resource; clear all. Article Type. Research And Reviews 3853; Peer Reviewed. Yes 3853; Publication Date. 1986. 2024. Apply. Last Year 704; Last 6 Months 385; Science Family of Journals The Journal of Remote Sensing, an Open Access journal published in association with AIR-CAS, promotes the theory, science, and technology of remote sensing, as well as interdisciplinary research within earth and information science. The Open Access journal Health Data Science, published in association with PKU, aims to advance the horizon of ... The New England Journal of Medicine The New England Journal of Medicine (NEJM) is a weekly general medical journal that publishes new medical research and review articles, and editorial opinion on a wide variety of topics of ... Frontiers Get the latest research updates, subscribe to our newsletter. Open access publisher of peer-reviewed scientific articles across the entire spectrum of academia. Research network for academics to stay up-to-date with the latest scientific publications, events, blogs and news. ACS Biomaterials Science & Engineering Journal Read current and featured research from the ACS Biomaterials Science & Engineering on ACS Publications, a trusted source for peer-reviewed journals. ... educational tools, free access to 50 ACS journal articles, and more! Join ACS. Researchers. Discover our blog, and stay current with your field. ... Based on recommendations from the scientific ... Scientific Journals Scientific Journals. AAAS publishes six respected peer-reviewed journals. Science, the premier global science weekly; Science Signaling, the leading journal of cell signaling and regulatory biology; Science Translational Medicine, integrating medicine, engineering and science to promote human health; Science Advances, an innovative and high ... Research articles Research articles. Filter By: Article Type. All. All; Article (208100) Conference Proceeding (56) Matters Arising (55) Registered Report (20) Year. All. All; 2024 (19155) 2023 (22198) 2022 (21835) 2021 (23320) ... About the journal. About Scientific Reports Contact Journal policies ... Browse Articles Research Highlights (4493) Scientific Correspondence (4028) Supplement to Nature (1556) Year. All. All; 2024 (2850) ... including the kind used to print certain high-profile journals. JSTOR Home Harness the power of visual materials—explore more than 3 million images now on JSTOR. Enhance your scholarly research with underground newspapers, magazines, and journals. Take your research further with Artstor's 3+ million images. Explore collections in the arts, sciences, and literature from the world's leading museums, archives, and ... Search Find the research you need | With 160+ million publication pages, 1+ million questions, and 25+ million researchers, this is where everyone can access science Find a journal Elsevier Journal Finder helps you find journals that could be best suited for publishing your scientific article. Journal Finder uses smart search technology and field-of-research specific vocabularies to match your paper's abstract to scientific journals. How do soundboard-trained dogs respond to human button presses? An Past research on interspecies communication has shown that animals can be trained to use Augmentative Interspecies Communication (AIC) devices, such as soundboards, to make simple requests of their caretakers. The recent uptake in AIC devices by hundreds of pet owners around the world offers a novel opportunity to investigate whether AIC is possible with owner-trained family dogs. To answer ... Open and free content on JSTOR and Artstor Journals. Explore our growing collection of Open Access journals. Early Journal Content, articles published prior to the last 95 years in the United States, or prior to the last 143 years if initially published internationally, are freely available to all. Even more content is available when you register to read - millions of articles from nearly 2,000 journals (IUCr) Accelerating imaging research at large-scale scientific 1. Introduction. X-ray imaging methods are widely used to investigate the structural properties of materials at the macro-, micro- and nano-scale, as well as magnetic domain structures, spatial distribution of the elements, chemical properties, etc. (Xu et al., 2016; Wang et al., 2020; Zhang et al., 2018, 2022; Suzuki et al., 2018).The wavelength of X-rays is significantly shorter than that of ... The effects of forging strategies on microstructures ... Fig.2 illustrates the microstructure of the as-cast Ti-55511 alloy. The OM image in Fig.2a reveals large grains, several millimeters in size, comprised of massive boundary α phase. Figs.2b and 2 c are SEM images showing the distribution of α phase both on the grain boundaries and within the grains. The boundary α phase with lamellar structure was precipitated near the grain boundaries, as ... Research and Reviews Single-molecule structural and kinetic studies across sequence space. by. Ivo Severins. Carolien Bastiaanssen. Sung Hyun Kim. Roy B. Simons. John van Noort. Chirlmin Joo. Science Vol. 385, NO. 6711 22 Aug 2024 : 898-904. Academic journal There are different types of peer-reviewed research journals; these specific publications are about food science.. An academic journal or scholarly journal is a periodical publication in which scholarship relating to a particular academic discipline is published. They serve as permanent and transparent forums for the presentation, scrutiny, and discussion of research. Top articles Top articles. Explore the most downloaded* papers from Scientific Reports in 2023. Featuring authors from around the world, these collections highlight valuable research from an international ... Organizing Your Social Sciences Research Paper How to Read (and Understand) a Social Science Journal Article. Inter-University Consortium for Political and Social Research (ISPSR), 2013. Shon, Phillip Chong Ho. How to Read Journal Articles in the Social Sciences: A Very Practical Guide for Students. 2nd edition. Thousand Oaks, CA: Sage, 2015; Lockhart, Tara, and Mary Soliday. Early science and colossal stone engineering in Menga, a Neolithic Here, we examine a great Neolithic engineering feat: the Menga dolmen, Iberia's largest megalithic monument. As listed by UNESCO, the Antequera megalithic site includes two natural formations, La Peña de los Enamorados and El Torcal karstic massif, and four major megalithic monuments: Menga, Viera, El Romeral, and the one recently discovered at Piedras Blancas, at the foot of La Peña de ... essay homework paper phd review speech study thesis