Scientific works
Master II thesis carried out at the University of Chambery
In collaboration with Vania Herbillon
Summary :
In the present study, we evaluated procedural and implicit learning in dyslexic children within the framework of cerebellar deficit theory. To do this, three groups of children (normal-reader subjects, phonological dyslexics and dyslexics without phonological disorder) took part in a serial reaction time task. Subjects performed 6 stimuli blocks of which the 2, 3, 4 and 6th were sequential blocks, the first and 5th were random blocks. The utility of the last random block (i.e., block 5) is to show that learning is not due to visuomotor training, but to learning the visuomotor sequence. Analysis of reaction times showed that children with phonological dyslexia learn a visuomotor sequence just as well as children with normal reading and children with surface dyslexia. Another statistical analysis taking into account the number of errors made by the subjects on random blocks versus sequential blocks showed a significant difference between blocks and between groups. In fact, surface dyslexics make a greater number of errors compared to normal readers. Thus, the learning of a visuomotor sequence takes place in dyslexics, but it is more difficult to establish in people with surface dyslexia. We propose that the two types of dyslexia are differentiated by the quality of their procedural learning of a visuomotor sequence, that is to say that dyslexics
phonology have better procedural learning of a visuomotor sequence than that of surface dyslexics. We propose to explain this difference by a deficit in the distribution of attentional resources allocated to a visuomotor stimulus existing in children with surface dyslexia.
An analysis of the profile of errors committed by surface dyslexics does not show any difference between the latter and normal-reading subjects. This indicates that the difference is not in the pattern of errors, but in the number of errors made by surface dyslexics. A sequence generation task allowed us to conclude that, unlike the two types of dyslexic children, normal-reading children became aware of the presence of a regularity in the appearance of the items in the sequence. On the contrary, the forced evocation of the order of the sequence as well as the recognition of the parts of the sequence show that learning remained implicit in all three groups. Our conclusion therefore consists of three points. The first is that children with phonological dyslexia learn a visuomotor sequence. This result does not agree with the cerebellar deficit hypothesis. The second, surface dyslexics learn a visuomotor sequence, but with many mistakes. This serial reaction time task is likely to provide additional information in order to differentiate dyslexics with a phonological disorder from dyslexics without a phonological disorder. Third, given the importance of the number of stimuli, the level of complexity of the visuomotor sequence and the type of presentation, future studies of procedural learning must consider each of these points before formulating conclusions about the absence or presence of procedural learning in dyslexic children. On a clinical level, dyslexic subjects not having realized the presence of a regularity in the presentation of the items of the sequence, would need to work on the clarification of the existing regularities in linguistic material.
Psychophysical analysis of the visual field: detection, identification, grouping effect and perceptual learning
Supervisor: Dr. Kenneth KNOBLAUCH, INSERM, BRON and Professor: Dennis LEVI, Eye Institute, Houston, TEXAS.
This work focused on three themes:
1- Peripheral vision: identification and detection of letters in three retinal eccentricities.
2- Decrease in visual acuity when the visual environment is loaded: effect of Grouping or Crowding phenomenon.
3- Incident perceptual learning related to the concept of Ahumada classification image (1996).
Summary :
The objective of this work is to understand how peripheral vision works. To do this, three studies were carried out. In the first, we measured the scaling factor required to achieve the same level of performance in detecting and identifying symmetrical letters with three eccentricities (2, 4 and 8ΚΏ). The results showed that a scaling factor achieves the same level of detection and identification performance across the peripheral visual field. However, an unexpected result was that the two tasks evolved identically across the periphery. The analysis of the errors of this experiment argues in favor of a superiority of the treatment of the tangential orientation in peripheral vision. A second experiment focused on the grouping effect (i.e., reduced visual acuity when a target is presented in a charged environment) that is measured by a stimulus composed of a target surrounded by distractors. Our results showed that the grouping effect extended over a distance equivalent to 20% of the eccentricity. We concluded that the clustering effect does not only depend on the similarity between target and border but also on the difference between the two. In order to open a new perspective for the study of perceptual learning in peripheral vision, we propose an innovative technique inspired by the classification image technique (Ahumada, 1996). Thus, the third series of experiments was devoted to learning a noisy stimulus. The observer had to detect the presence or absence of a signal of which he did not know the form. The results showed that the observer could learn a signal of which he did not know the form, this learning required a multi-modal feedback (visual and auditory). In conclusion, peripheral vision is not a reduced version of foveal vision. It has its own characteristics which allow it to process the flow of information that comes from the environment.
Modeling of a psychometric function in R. Rosa Yssaad Fesselier, Kenneth Knoblauch. Behavioral research method. February 2006, 38 (1) 28-41.
Goals :
'Model psychometric functions with R.'
We used a statistical computation on the R software to obtain the estimate of the fit of a psychometric function. We fitted the data using a nonlinear regression model.
Abstract :
We demonstrate some procedures in the R statistical computational environment to obtain maximum likelihood estimates of parameters of a psychometric function by fitting a generalized nonlinear regression model to the data. A feature to fit a linear model to the threshold (or other) parameters of multiple psychometric functions simultaneously provides a powerful tool for testing hypotheses on the data and, potentially, for reducing the number of parameters needed to describe them. Finally, we illustrate procedures for treating a parameter as a random effect that would allow a simplified approach to modeling stimulus-independent variability due to factors such as failures or differences between observers. These tools will facilitate a more comprehensive and explicit approach to modeling psychometric data.
Site address:
Intellectual and visuospatial evaluation in long-term pediatric liver transplantation for biliary atresia. Rosa Yssaad-Feselier et al. Transplantation. May 15, 2009; 87 (9).
The majority of parents observed difficulty concentrating and slowing down visuomotor tasks in their transplanted child. A group of children aged between 7 and 13 years old who had received liver transplants for an average of ten years participated in this study.
The results showed that these children have normal intelligence, but have difficulty in two areas: concentration (a low working memory index) and visuospatial.
site address: