Dr Sylvain Sirois

Photograph of Sylvain Sirois

Honorary lecturer


Infant Habituation

Habituation provides researchers with a unique window into the infant's mind. Infants can't speak, can't fill out questionnaires, and aren't all that eager to follow instructions, however simple. What they do quite well, though, is become bored over repeated presentations of stimuli (i.e., they habituate). Moreover, they can show renewed interest when you change the stimuli (i.e., they dishabituate). Using this indirect approach, researchers have investigated infants' perceptual, categorical, and cognitive abilities (by examining what they seemingly distinguish through renewed interest). This area of research is currently quite active, and is associated with possibly the most heated debate in psychology regarding the cognitive abilities of young infants. My work in this area is two-fold. Firstly, I am working on a neural network model of habituation that attempts to bridge the gap between neural circuits believed to be involved in habituation and the various behavioural manifestations of habituation. Secondly, I am in the process of setting up an infancy lab to test some of the unique predictions of the model. This combination of computational and empirical work should ultimately provide an accurate characterisation of habituation as a process. It will then be possible to address the issue of infant cognition from solid ground (rather than confuse it with simple learning processes, as is currently the case, hence the heated debate). Visit the BabyLab's web page here.

Learning and Development

Because we use these two distinct words, we assume that learning and development are two different things. Yet when the two processes are described, it is usually in terms of outcomes rather than mechanisms. Thus, development is used to describe behaviour changes that are broad in application, that take place over a long period, that are rather universal, and that take place mostly in children. Learning, conversely, is used to describe changes that are specific, quick, idiosyncratic, and that take place in children and adults alike. Upon closer scrutiny, most theoretical work in the area of cognitive change fails to make a distinction between learning and development as processes. My work in this area is to formulate a viable, formal distinction between both processes. This involves a neural network framework whereby learning and development are implemented as qualitatively distinct mechanisms, making unique and complementary contributions to cognitive change. As with my work on habituation, the goal is to formulate this framework within the constraints of brain circuit properties. So far, this area of research was instrumental in clarifying vague but central ideas in Piagetian theory.

Shift Learning

This is an area of research where a lot has been said about young children's abilities to learn categories. Shift learning tasks involve the acquisition of arbitrary categories based (usually) on perceptual features such as shape, colour, etc. When performance is high, a new category is introduced and thus people must shift their responses. It turns out that preschoolers perform differently than older children and adults on a variety of tasks, which suggested that their categories are different and develop between the ages of roughly 4 to 10. However, if you provide preschoolers with 20 or so additional learning trials, they perform as adults. The suggestion that they might "develop" in 20 trials rather than several years is likely preposterous. It may be, therefore, that their categories are not that different. My work in this area has been to formulate a theoretical model of shift learning based on the assumption that the main source of difference between preschoolers and adults is rehearsal ability. Whereas much research had shown how preschoolers could perform as adults, I've shown that adults can perform as preschoolers if you block rehearsal. This strengthens the case that categories are possibly quite similar in adults and young children, and that other, general factors affect performance.

The Development of Information Integration

We take for granted that a quantity of water poured from a tall, thin glass into a short, broad glass remains the same quantity of water. (If we're really picky, we could say that it isn't because there's a thin film of water remaining in the first glass, but such fussiness is usually frowned upon.) We also take for granted that objects hidden in one of many distinct boxes are even easier to find if the boxes aren't shuffled about. It also seems quite reasonable that when different weights are on both arms of a scale, only one side will go down, and only one would go up. Yet in some conditions children believe that a quantity of water changes when the shape of the container is changed. They can also find it easier to retrieve a toy from one of several boxes when the correct one has been moved, compared to when it is at the initial place. And children are happy to report that in some instances both arms of a balance will go down or up. What is common to all these tasks and that can cause problems for children is that more than one source of information must be integrated in order to represent adequately the problem at hand. In the case of water, volume equals height times width times depth, and thus focusing on only one dimension is misleading. In the case of hiding boxes, spatial and non-spatial information about boxes must be separated (so that when boxes are shuffled, the target box can be identified independently of its original location). Trouble arises when children come to view box identity and original location as mutually exclusive. Finally, with the balance task, the fact that both sides have large weights doesn't imply that both will go down. The relative torque on each side of the arm predicts which (if any) of the two sides goes down. My work in this area involves charting the emergence of information integration, as well as looking at the factors that influence children's errors. For example, children's usual height bias (i.e. tall glasses hold more) in the water conservation task is replaced by a width bias if liquids travel horizontally rather than vertically (through cunning computer graphics).


  • PSYC10101 Statistics and Research Design
  • PSYC60301 Advanced Statistics in Psychology


After obtaining my BSc in psychology in 1995, I enrolled for a PhD at McGill University, in Montreal, which I completed in 2000. I then spent two years as an associate research fellow at the Centre for Brain and Cognitive Development, Birkbeck College, London. I have been a lecturer with the Department of Psychology (now School of Psychological Sciences), The University of Manchester, since August 2002. I am a co-author and a co-editor of the Neuroconstructivism books.

Collaborators and affiliated staff


  • Prof. Denis Mareschal
    Centre for Brain and Cognitive Development, Birkbeck University of London
  • Prof. Thomas R. Shultz
    Department of Psychology, McGill University, Montreal
  • Profs. Claude Dumas and Henry Markovits
    Department of Psychology, Université du Québec à Montréal, Montreal

Selected publications


  • Sirois S, Iain Jackson. (2007). Social cognition in infancy: a critical review of research on higher-order abilities. European Journal of Developmental Psychology, 4, eScholarID:1d13700


  • Sirois S, Thomas R. Shultz. (2006). Preschoolers out of adults: Discriminative learning with a cognitive load. Quarterly Journal of Experimental Psychology, eScholarID:1d11060 | DOI:10.1080/17470210500246269


  • Sirois S, Mareschal D. (2004). An interacting systems model of infant habituation. Journal of Cognitive Neuroscience, 16(8), 1352-1362. eScholarID:1d8328 | DOI:10.1162/0898929042304778


View all publications