Nayla Sokhn

iMap4: An open source toolbox for the statistical fixation mapping of eye movement data with linear mixed modeling

A major challenge in modern eye movement research is to statistically map where observers are looking, by isolating the significant differences between groups and conditions. As compared to the signals from contemporary neuroscience measures, such as magneto/electroencephalography and functional magnetic resonance imaging, eye movement data are sparser, with much larger variations in space across trials and participants. As a result, the implementation of a conventional linear modeling approach on two-dimensional fixation distributions often returns unstable estimations and underpowered results, leaving this statistical problem unresolved (Liversedge, Gilchrist, & Everling, 2011). Here, we present a new version of the iMap toolbox (Caldara & Miellet, 2011) that tackles this issue by implementing a statistical framework comparable to those developed in state-of-the-art neuroimaging data-processing toolboxes. iMap4 uses univariate, pixel-wise linear mixed models on smoothed fixation data, with the flexibility of coding for multiple between- and within-subjects comparisons and performing all possible linear contrasts for the fixed effects (main effects, interactions, etc.). Importantly, we also introduced novel nonparametric tests based on resampling, to assess statistical significance. Finally, we validated this approach by using both experimental and Monte Carlo simulation data. iMap4 is a freely available MATLAB open source toolbox for the statistical fixation mapping of eye movement data, with a user-friendly interface providing straightforward, easy-to-interpret statistical graphical outputs. iMap4 matches the standards of robust statistical neuroimaging methods and represents an important step in the data-driven processing of eye movement fixation data, an important field of vision sciences.

iMap 4: An Open Source Toolbox for the Statistical Fixation Mapping of Eye Movement data with Linear Mixed Modeling

A major challenge in modern eye movement research is to statistically map where observers are looking at, as well as isolating statistical significant differences between groups and conditions. Compared to signals of contemporary neuroscience measures, such as M/EEG and fMRI, eye movement data are sparse with much larger variations across trials and participants. As a result, the implementation of a conventional Hierarchical Linear Model approach on two-dimensional fixation distributions often returns unstable estimations and underpowered results, leaving this statistical problem unresolved. Here, we tackled this issue by using the statistical framework implemented in diverse state-of-the-art neuroimaging data processing toolboxes: Statistical Parametric Mapping (SPM), Fieldtrip and LIMO EEG. We first estimated the mean individual fixation maps per condition by using trimmean to account for the sparseness and the high variations of fixation data. We then applied a univariate, pixel-wise linear mixed model (LMM) on the smoothed fixation data with each subject as a random effect, which offers the flexibility to code for multiple between- and within- subject comparisons. After this step, our approach allows to perform all the possible linear contrasts for the fixed effects (main effects, interactions, etc.). Importantly, we also introduced a novel spatial cluster test based on bootstrapping to assess the statistical significance of the linear contrasts. Finally, we validated this approach by using both experimental and computer simulation data with a Monte Carlo approach. iMap 4 is a freely available MATLAB open source toolbox for the statistical fixation mapping of eye movement data, with a user-friendly interface providing straightforward, easy to interpret statistical graphical outputs and matching the standards in robust statistical neuroimaging methods. iMap 4 represents a major step in the processing of eye movement fixation data, paving the way to a routine use of robust data-driven analyses in this important field of vision sciences. Meeting abstract presented at VSS 2015.

Identification of Chordless Cycles in Ecological Networks

In the last few years the studies on complex networks have gained extensive research interests. Significant impacts are made by these studies on a wide range of different areas including social networks, technology networks, biological networks and others. Motivated by understanding the structure of ecological networks we introduce in this paper a new algorithm for enumerating all chordless cycles. The proposed algorithm is a recursive one based on the depth-first search.

A left eye bias for female faces

We investigated whether men and women use different scanning strategies to extract the visual information relevant for the recognition of female and male faces. We used an old-new recognition task during which observers were asked to identify previously learned faces. Data from two groups of observers (male, female) revealed a more accurate, but not faster, recognition of male face stimuli. Interestingly, the fixation maps revealed a consistent left gaze bias for female face stimuli, regardless of the gender of the observer. Altogether, our data show that humans deploy distinct and flexible visual sample strategies to process faces.

Structural Network Properties of Niche-Overlap Graphs

The structure of networks has always been interesting for researchers. Investigating their unique architecture allows to capture insights and to understand the function and evolution of these complex systems. Ecological networks such as food-webs and niche-overlap graphs are considered as complex systems. The main purpose of this work is to compare the topology of 15 real niche-overlap graphs with random ones. Five measures are treated in this study: (1) the clustering coefficient, (2) the between ness centrality, (3) the assortativity coefficient, (4) the modularity and (5) the number of chord less cycles. Significant differences between real and random networks are observed. Firstly, we show that niche-overlap graphs display a higher clustering and a higher modularity compared to random networks. Moreover we find that random networks have barely nodes that belong to a unique sub graph (i.e. between ness centrality equal to 0) and highlight the presence of a small number of chord less cycles compared to real networks. These analyses may provide new insights in the structure of these real niche-overlap graphs and may give important implications on the functional organization of species competing for some resources and on the dynamics of these systems.

A decisional space account of saccadic reaction times towards personally familiar faces

Manual and saccadic reaction times (SRTs) have been used to determine the minimum time required for different types of visual categorizations. Such studies have demonstrated that faces can be detected within natural scenes within as little as 100ms (Crouzet, Kirchner & Thorpe, 2010), while increasingly complex decisions require longer processing times (Besson, Barragan-Jason, Thorpe, Fabre-Thorpe, Puma et al., 2017). Following the notion that facial representations stored in memory facilitate perceptual processing (Ramon & Gobbini, 2018), a recent study reported 180ms as the fastest speed at which “familiar face detection” based on expressed choice saccades (Visconti di Ollegio Castello & Gobbini, 2015). At first glance, these findings seem incompatible with the earliest neural markers of familiarity reported in electrophysiological studies (Barragan-Jason, Cauchoix & Barbeau, 2015; Caharel, Ramon & Rossion, 2014; Huang, Wu, Hu, Wang, Ding & Qu et al., 2017), which should temporally precede any overtly observed behavioral (oculomotor or manual) categorization. Here, we reason that this apparent discrepancy could be accounted for in terms of decisional space constraints, which modulate both manual RTs observed for different levels of visual processing (Besson et al., 2017), as well as saccadic RTs (SRTs) in both healthy observers and neurological patients (Ramon, in press; Ramon, Sokhn, Lao & Caldara, in press). In the present study, over 70 observers completed three different SRT experiments in which decisional space was manipulated through task demands and stimulus probability. Subjects performed a gender categorization task, or one of two familiar face “recognition” tasks, which differed with respect to the number of personally familiar identities presented (3 vs. 7). We observe an inverse relationship between visual categorization proficiency and decisional space. Observers were most accurate for categorization of gender, which could be achieved in as little as 140ms. Categorization of highly predictable targets was more error-prone and required an additional ~100ms processing time. Our findings add to increasing evidence that pre-activation of identity-information can modulate early visual processing in a top-down manner. They also emphasize the importance of considering procedural aspects as well as terminology when aiming to characterize cognitive processes.

Decisional space determines saccadic reaction times in healthy observers and acquired prosopagnosia

ABSTRACT Determining the familiarity and identity of a face have been considered as independent processes. Covert face recognition in cases of acquired prosopagnosia, as well as rapid detection of familiarity have been taken to support this view. We tested P.S. a well-described case of acquired prosopagnosia, and two healthy controls (her sister and daughter) in two saccadic reaction time (SRT) experiments. Stimuli depicted their family members and well-matched unfamiliar distractors in the context of binary gender, or familiarity decisions. Observers’ minimum SRTs were estimated with Bayesian approaches. For gender decisions, P.S. and her daughter achieved sufficient performance, but displayed different SRT distributions. For familiarity decisions, her daughter exhibited above chance level performance and minimum SRTs corresponding to those reported previously in healthy observers, while P.S. performed at chance. These findings extend previous observations, indicating that decisional space determines performance in both the intact and impaired face processing system.

Structure comparison of binary and weighted niche-overlap graphs

In ecological networks, niche-overlap graphs are considered as complex systems. They represent the competition between two predators that share common resources. The purpose of this paper is to investigate the structural properties of these graphs considered as weighted networks and compare their measures with the ones calculated for the binary networks. To conduct this study, we select four classical network measures : the degree of nodes, the clustering coefficient, the assortativity, and the betweenness centrality. These measures were used to analyse different type of networks such as social networks, biological networks, world wide web, etc. Interestingly, we identify significant differences between the structure of the binary and the weighted niche-overlap graphs. This study indicates that weight information reveals different features that may provide other implications on the dynamics of these networks.