Benoît Lemaire
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Benoît Lemaire has a total of 70 co-authors. Here are the top co-authors by number of shared publications:
Anne Guérin-Dugué
Univ. Grenoble Alpes, CNRS, Grenoble INP, GIPSA-lab, 38000, Grenoble, France
6 shared publicationsSophie Portrat
5 shared publications- Gaën Plancher
Laboratoire d’Étude des Mécanismes Cognitifs (EMC – EA 3082), Université Lumière Lyon 2, Lyon, France
4 shared publications - Sophie Portrat
Laboratoire de Psychologie et Neurocognition (LPNC, CNRS 5105), Université Grenoble Alpes, Université Savoie Mont Blanc, Grenoble, France
4 shared publications - Anne Guérin-Dugué
University Grenoble Alpes, CNRS, Grenoble INP+, GIPSA-Lab, (+)Institute of Engineering University Grenoble Alpes, 38000 Grenoble, France
3 shared publications - Catherine Thevenot
Institute of Psychology University of Lausanne
3 shared publications
Latest Publications
- 21 April 2026
- Published by Wiley in Cognitive Science
- Vol. 50 (4) , e70207
A Computational Model of Basic Addition Solving
This article presents a computational learning model in which procedural execution and memory retrieval codevelop, using simple arithmetic, which provides a particularly well‐controlled domain for investigating this issue. In single‐digit addition learning, strategies employed initially rely on counting due to the lack of stored answers in memory. Over time, associations between problems and their solutions are strengthened. The model accounts for this learning process by dynamically selecting between counting and memory retrieval, based on their expected duration. It also introduces a mechanism for accelerating counting throuSUPPLEMgh repeated practice along the mental sequence. The model was first tested on data collected from adults learning to solve alphabet arithmetic problems over a 3‐week experiment. It successfully replicated the empirical finding that larger problems are memorized earlier than smaller ones. A second simulation was conducted using data from an experiment manipulating problem structure: participants were trained on either contiguous (A+…, B+…, C+…) or noncontiguous (A+…, C+…, E+…) sequences. This variation affected the transition between strategies: participants in the noncontiguous condition showed a greater tendency to rely on retrieval, as the practice of moving from one letter to the next differed. The model also reproduced this pattern. Overall, the results suggest that no single strategy dominates at the end of learning; rather, counting and retrieval coexist, depending on problem size and structure. This model is, to our knowledge, the only one to incorporate a counting acceleration mechanism in line with the automated counting theory and memory retrieval.
- 21 April 2026
- Published by Wiley in Cognitive Science
- Vol. 50 (4) , e70207
A Computational Model of Basic Addition Solving
This article presents a computational learning model in which procedural execution and memory retrieval codevelop, using simple arithmetic, which provides a particularly well‐controlled domain for investigating this issue. In single‐digit addition learning, strategies employed initially rely on counting due to the lack of stored answers in memory. Over time, associations between problems and their solutions are strengthened. The model accounts for this learning process by dynamically selecting between counting and memory retrieval, based on their expected duration. It also introduces a mechanism for accelerating counting throuSUPPLEMgh repeated practice along the mental sequence. The model was first tested on data collected from adults learning to solve alphabet arithmetic problems over a 3‐week experiment. It successfully replicated the empirical finding that larger problems are memorized earlier than smaller ones. A second simulation was conducted using data from an experiment manipulating problem structure: participants were trained on either contiguous (A+…, B+…, C+…) or noncontiguous (A+…, C+…, E+…) sequences. This variation affected the transition between strategies: participants in the noncontiguous condition showed a greater tendency to rely on retrieval, as the practice of moving from one letter to the next differed. The model also reproduced this pattern. Overall, the results suggest that no single strategy dominates at the end of learning; rather, counting and retrieval coexist, depending on problem size and structure. This model is, to our knowledge, the only one to incorporate a counting acceleration mechanism in line with the automated counting theory and memory retrieval.
- 10 February 2025
- Published by Taylor & Francis in Journal of Cognitive Psychology
- Vol. 37 (3) , 249-266
How learning material shapes learning strategies in an alphabet arithmetic task
- 10 February 2025
- Published by Taylor & Francis in Journal of Cognitive Psychology
- Vol. 37 (3) , 249-266
How learning material shapes learning strategies in an alphabet arithmetic task
- 12 August 2024
- Published by Springer Nature in Memory & Cognition
- Vol. 52 (8) , 2022-2052
Does the extension of free time trigger spontaneous elaborative strategies in working memory?
Elaboration has emerged as a potential maintenance mechanism involved in the substantial contribution of long-term memory (LTM) to working memory (WM) perf
- 12 August 2024
- Published by Springer Nature in Memory & Cognition
- Vol. 52 (8) , 2022-2052
Does the extension of free time trigger spontaneous elaborative strategies in working memory?
Elaboration has emerged as a potential maintenance mechanism involved in the substantial contribution of long-term memory (LTM) to working memory (WM) perf
- 1 March 2024
- Published by American Psychological Association (APA) in Journal of Experimental Psychology: Learning, Memory, and Cognition
- Vol. 50 (3) , 418-434
Learning basic arithmetic: A comparison between rote and procedural learning based on an artificial sequence.
It is commonly accepted that repeatedly using mental procedures results in a transition to memory retrieval, but the determinant of this process is still unclear. In a 3-week experiment, we compared two different learning situations involving basic additions, one based on counting and the other based on arithmetic fact memorization. Two groups of participants learned to verify additions such as "G + 2 = Q?" built on an artificial sequence (e.g., "XGRQD…"). The first group learned the sequence beforehand and could therefore count to solve the problems, whereas the second group was not aware of the sequence and had to learn the equations by rote. With practice, solution times of both groups reached a plateau, indicating a certain level of automatization. However, a more fine-grained comparison indicated that participants relied on fundamentally different learning mechanisms. In the counting condition, most participants showed a persistent linear effect of the numerical operand on solution times, suggesting that fluency was reached through an acceleration of counting procedures. However, some participants began memorizing the problems involving the largest addends: Their solution times were very similar to those of participants in the rote learning group, suggesting that they resulted from a memory retrieval process. These findings show that repeated mental procedures do not systematically lead to memory retrieval but that fluency can also be reached through the acceleration of these procedures. Moreover, these results challenge associationist models, which cannot currently predict that the process of memorization begins with problems involving the largest addends. (PsycInfo Database Record (c) 2023 APA, all rights reserved).
- 1 March 2024
- Published by American Psychological Association (APA) in Journal of Experimental Psychology: Learning, Memory, and Cognition
- Vol. 50 (3) , 418-434
Learning basic arithmetic: A comparison between rote and procedural learning based on an artificial sequence.
It is commonly accepted that repeatedly using mental procedures results in a transition to memory retrieval, but the determinant of this process is still unclear. In a 3-week experiment, we compared two different learning situations involving basic additions, one based on counting and the other based on arithmetic fact memorization. Two groups of participants learned to verify additions such as "G + 2 = Q?" built on an artificial sequence (e.g., "XGRQD…"). The first group learned the sequence beforehand and could therefore count to solve the problems, whereas the second group was not aware of the sequence and had to learn the equations by rote. With practice, solution times of both groups reached a plateau, indicating a certain level of automatization. However, a more fine-grained comparison indicated that participants relied on fundamentally different learning mechanisms. In the counting condition, most participants showed a persistent linear effect of the numerical operand on solution times, suggesting that fluency was reached through an acceleration of counting procedures. However, some participants began memorizing the problems involving the largest addends: Their solution times were very similar to those of participants in the rote learning group, suggesting that they resulted from a memory retrieval process. These findings show that repeated mental procedures do not systematically lead to memory retrieval but that fluency can also be reached through the acceleration of these procedures. Moreover, these results challenge associationist models, which cannot currently predict that the process of memorization begins with problems involving the largest addends. (PsycInfo Database Record (c) 2023 APA, all rights reserved).
- 1 April 2021
- Published by Elsevier in Journal of Memory and Language
- Vol. 117
How does semantic knowledge impact working memory maintenance? Computational and behavioral investigations
- 1 April 2021
- Published by Elsevier in Journal of Memory and Language
- Vol. 117
How does semantic knowledge impact working memory maintenance? Computational and behavioral investigations
- 9 May 2018
- Published by Wiley in Annals of the New York Academy of Sciences
- Vol. 1424 (1) , 149-160
What makes working memory traces stable over time?
Understanding the factors that make working memory (WM) traces stable over time is important because WM is the keystone of general cognitive achievement. Two views of attentional WM maintenance have been suggested to account for the long‐term retention of WM information. First, the distractors in a WM task are thought to foster the creation of episodic memory cues through covert retrieval. Second, the cognitive load (CL) of the distractors is thought to vary long‐term memory instantiation. In this study, we propose an additional parsimonious perspective: the total time during which each trace is under the control of attention in WM is the key to long‐term retention. Participants performed a complex span task in which the CL and number of distractors were orthogonally manipulated, and thereafter the participants performed a delayed recall test. Similar to previous findings, the results showed effects of the number of distractors and of CL on delayed recall. Our results went further, however, by showing a non‐linear relationship between delayed recall performance and the free time accumulated between encoding and immediate recall. The role of time in episodic memory performance and the underlying WM maintenance mechanism are discussed.
- 9 May 2018
- Published by Wiley in Annals of the New York Academy of Sciences
- Vol. 1424 (1) , 149-160
What makes working memory traces stable over time?
Understanding the factors that make working memory (WM) traces stable over time is important because WM is the keystone of general cognitive achievement. Two views of attentional WM maintenance have been suggested to account for the long‐term retention of WM information. First, the distractors in a WM task are thought to foster the creation of episodic memory cues through covert retrieval. Second, the cognitive load (CL) of the distractors is thought to vary long‐term memory instantiation. In this study, we propose an additional parsimonious perspective: the total time during which each trace is under the control of attention in WM is the key to long‐term retention. Participants performed a complex span task in which the CL and number of distractors were orthogonally manipulated, and thereafter the participants performed a delayed recall test. Similar to previous findings, the results showed effects of the number of distractors and of CL on delayed recall. Our results went further, however, by showing a non‐linear relationship between delayed recall performance and the free time accumulated between encoding and immediate recall. The role of time in episodic memory performance and the underlying WM maintenance mechanism are discussed.
- Open Access
- 3 April 2018
- Published by Frontiers Media SA in Frontiers in Psychology
- Vol. 9, 416
A Computational Model of Working Memory Integrating Time-Based Decay and Interference
There is still a strong debate in the working memory literature about the cause of forgetting, with many articles providing evidence for the existence of temporal decay and as many publications providing evidence compatible with interference being the only mechanism involved in forgetting. In order to reconcile the two views, this article describes TBRS∗-I (for Time-Based Resource-Sharing∗-Interference), a computational model of working memory which incorporates an interference-based mechanism to the decay-based implementation TBRS∗ within the TBRS theoretical framework. At encoding, memoranda are associated to their context, namely their position in the list. Temporal decay decreases the strength of these associations, but a refreshing process may reactivate it during free time. Distractors may alter the distributed representation of memoranda but refreshing can restore them based on the long-term memory representations. Refreshing is therefore twofold: reactivation plus restoration, each one counteracting the detrimental time-based and interference-based decays, respectively. Two types of interference are implemented: interference by confusion which depends on the degree of overlap between memoranda and distractors and interference by superposition which depends on the similarity between them. TBRS∗-I was tested on six benchmark findings on retention-interval and distractor-processing effects by means of millions of simulations testing the effects of seven factors on memory performance: the number of memoranda, the duration of distractor attentional capture, the duration of free time, the number of distractors, the amount of overlap between memoranda and distractors, the similarity between memoranda and distractors and the homogeneity of distractors (all identical or all distinct). TBRS∗-I replicated classical effects and proved to be a suitable hybrid model integrating both interference and time-based decay. The article also discusses the compatibility of TBRS∗-I with a unitary or dual view of memory and the issue of integrating time and interference in a single model. Computer codes and data are available at https://osf.io/65sna/.
- 3 April 2018
- Published by Frontiers Media SA in Frontiers in Psychology
- Vol. 9, 416
Open AccessA Computational Model of Working Memory Integrating Time-Based Decay and Interference
There is still a strong debate in the working memory literature about the cause of forgetting, with many articles providing evidence for the existence of temporal decay and as many publications providing evidence compatible with interference being the only mechanism involved in forgetting. In order to reconcile the two views, this article describes TBRS∗-I (for Time-Based Resource-Sharing∗-Interference), a computational model of working memory which incorporates an interference-based mechanism to the decay-based implementation TBRS∗ within the TBRS theoretical framework. At encoding, memoranda are associated to their context, namely their position in the list. Temporal decay decreases the strength of these associations, but a refreshing process may reactivate it during free time. Distractors may alter the distributed representation of memoranda but refreshing can restore them based on the long-term memory representations. Refreshing is therefore twofold: reactivation plus restoration, each one counteracting the detrimental time-based and interference-based decays, respectively. Two types of interference are implemented: interference by confusion which depends on the degree of overlap between memoranda and distractors and interference by superposition which depends on the similarity between them. TBRS∗-I was tested on six benchmark findings on retention-interval and distractor-processing effects by means of millions of simulations testing the effects of seven factors on memory performance: the number of memoranda, the duration of distractor attentional capture, the duration of free time, the number of distractors, the amount of overlap between memoranda and distractors, the similarity between memoranda and distractors and the homogeneity of distractors (all identical or all distinct). TBRS∗-I replicated classical effects and proved to be a suitable hybrid model integrating both interference and time-based decay. The article also discusses the compatibility of TBRS∗-I with a unitary or dual view of memory and the issue of integrating time and interference in a single model. Computer codes and data are available at https://osf.io/65sna/.
- 15 February 2018
- Published by Springer Nature in Brain Topography
- Vol. 31 (4) , 640-660
An Eye Fixation-Related Potential Study in Two Reading Tasks: Reading to Memorize and Reading to Make a Decision
We investigated how two different reading tasks, namely reading to memorize [Read & Memorize (RM)] and reading to decide whether a text was relevant to a given topic [Read & Decide (RD)], modulated both eye movements (EM) and brain activity. To this end, we set up an ecological paradigm using the eye fixation-related potentials (EFRP) technique, in which participants freely moved their eyes to process short paragraphs, while their electroencephalography (EEG) activity was recorded in synchronization with their EM. A general linear model was used to estimate at best EFRP, taking account of the overlap between adjacent potentials, and more precisely with the potential elicited at text onset, as well as saccadic potentials. Our results showed that EM patterns were top-down modulated by different task demands. More interestingly, in both tasks, we observed slow-wave potentials that gradually increased across the first eye fixations. These slow waves were larger in the RD task than in the RM task, specifically over the left hemisphere. These results suggest that the decision-making process during reading in the RD task engendered a greater memory load in working memory than that generated in a classic reading task. The significance of these findings is discussed in the light of recent theories and models of working memory processing.
- 15 February 2018
- Published by Springer Nature in Brain Topography
- Vol. 31 (4) , 640-660
An Eye Fixation-Related Potential Study in Two Reading Tasks: Reading to Memorize and Reading to Make a Decision
We investigated how two different reading tasks, namely reading to memorize [Read & Memorize (RM)] and reading to decide whether a text was relevant to a given topic [Read & Decide (RD)], modulated both eye movements (EM) and brain activity. To this end, we set up an ecological paradigm using the eye fixation-related potentials (EFRP) technique, in which participants freely moved their eyes to process short paragraphs, while their electroencephalography (EEG) activity was recorded in synchronization with their EM. A general linear model was used to estimate at best EFRP, taking account of the overlap between adjacent potentials, and more precisely with the potential elicited at text onset, as well as saccadic potentials. Our results showed that EM patterns were top-down modulated by different task demands. More interestingly, in both tasks, we observed slow-wave potentials that gradually increased across the first eye fixations. These slow waves were larger in the RD task than in the RM task, specifically over the left hemisphere. These results suggest that the decision-making process during reading in the RD task engendered a greater memory load in working memory than that generated in a classic reading task. The significance of these findings is discussed in the light of recent theories and models of working memory processing.
- 26 September 2017
- Published by Taylor & Francis in Experimental Aging Research
- Vol. 43 (5) , 409-429
Under Which Conditions Can Older Participants Maintain Information In Working Memory?
Background/Study Context: Age-related deficits in working memory (WM) are commonly described in the literature. However, age-related deficits in WM maintenance mechanisms have received less attention. Based on recent models of WM, the present study aims at determining the conditions under which older adults are able to maintain information. More precisely, we wondered whether the cognitive load effect, taken as evidence for active maintenance in WM, is observed in normal aging and whether it interacts with the level of interference of a concurrent task. Methods: Young (mean age = 21.62; SD = 2.51) and healthy older (mean age = 71.92; SD = 5.18) participants performed a complex WM task. They had to remember five images while reading three words presented after each image. We compared trials in which every word was new (inducing high interference) with trials where words were repeated (inducing low interference). The pace at which the reading task was performed was either fast or slow, resulting in a high or a low cognitive load. Results: As suggested in the literature, young participants presented better performance at slow pace compared with fast pace but were not influenced by interference. Older participants also performed better at slow pace but only when interference was low. Interestingly, the older population showed negative correlations between slow-pace trials and switching abilities. Finally, although computational simulations with time-based resource sharing* (TBRS*) provide a good fit for young adult performance, several parameters had to be adjusted to fit the older participants’ performance, including duration of trace refreshing. Conclusion: Our findings suggest that a decrease in WM performance with aging can be explained by a difficulty in taking advantage of WM maintenance opportunities, especially in conditions of high interference. The computational investigations are consistent with this interpretation given that the parameters to be adjusted involve maintenance in WM. Finally, the computational approach seems a relevant way to address causes of forgetting in aging.
- 26 September 2017
- Published by Taylor & Francis in Experimental Aging Research
- Vol. 43 (5) , 409-429
Under Which Conditions Can Older Participants Maintain Information In Working Memory?
Background/Study Context: Age-related deficits in working memory (WM) are commonly described in the literature. However, age-related deficits in WM maintenance mechanisms have received less attention. Based on recent models of WM, the present study aims at determining the conditions under which older adults are able to maintain information. More precisely, we wondered whether the cognitive load effect, taken as evidence for active maintenance in WM, is observed in normal aging and whether it interacts with the level of interference of a concurrent task. Methods: Young (mean age = 21.62; SD = 2.51) and healthy older (mean age = 71.92; SD = 5.18) participants performed a complex WM task. They had to remember five images while reading three words presented after each image. We compared trials in which every word was new (inducing high interference) with trials where words were repeated (inducing low interference). The pace at which the reading task was performed was either fast or slow, resulting in a high or a low cognitive load. Results: As suggested in the literature, young participants presented better performance at slow pace compared with fast pace but were not influenced by interference. Older participants also performed better at slow pace but only when interference was low. Interestingly, the older population showed negative correlations between slow-pace trials and switching abilities. Finally, although computational simulations with time-based resource sharing* (TBRS*) provide a good fit for young adult performance, several parameters had to be adjusted to fit the older participants’ performance, including duration of trace refreshing. Conclusion: Our findings suggest that a decrease in WM performance with aging can be explained by a difficulty in taking advantage of WM maintenance opportunities, especially in conditions of high interference. The computational investigations are consistent with this interpretation given that the parameters to be adjusted involve maintenance in WM. Finally, the computational approach seems a relevant way to address causes of forgetting in aging.