Community Research and Development Information Service - CORDIS


facessvep Report Summary

Project ID: 284025
Funded under: FP7-IDEAS-ERC
Country: Belgium


The overarching goal of the ERC research project was to understand how humans perceive and recognize people by their face, an outstanding issue in vision science, neuroscience and artificial intelligence. To reach this objective, we introduced a novel approach in the field of face perception, and high-level vision in general: we present face or object continuously for one to two minutes, at a rapid fixed rate, for instance 6 times by second (6 Hz), and we measure the electrical response of the brain exactly at that rate. Throughout the project, we made substantial progress in understanding the approach at a methodological level as well as answering important issues in face perception research. For instance, we have defined for the first time the optimal rate at which your brain can recognize individual faces (i.e. how many faces can you discriminate in one second ?). The answer is 6 Hz (i.e. 6 faces/ second; Alonso-Prieto et al., 2013; Gentile & Rossion, 2014).
We showed that the sensitivity to individual faces as identified by this approach is specific to upright faces and faces with typical contrast polarity (i.e., eyes/eyebrows darker than skin, Rossion et al., 2012; Liu-Shuang et al., 2014). We separated the contribution of the shape vs. the colour and texture information in the contribution to this response, showing for the first time that the combination of the two sources of information gives a face its identity, more than the sum of the 2 properties presented separately (Dzhelyova & Rossion, 2014). This latter work has been successful thanks to the development of a fast “oddball” periodic stimulation paradigm (Liu-Shuang et al., 2014; 2016) in which we present the same stimulus at a fixed rate and we interrupt it by other stimuli at a lower fixed rate (e.g., AAAABAAAACAA...). This development has been so successful that the paradigm is now used for a majority of our studies, and has been borrowed by other laboratories. This new methodology has even been used to detect face-selective responses to widely variable images of faces over the right hemisphere in infants as early as 4 months old (de Heering & Rossion, 2015)
The neural basis of the periodic responses has been identified with the recording of intracerebral potentials in large groups of patients, providing the first evidence for hemispheric and regional specialization of visual category-selectivity in the human brain with direct measures of brain activity. Finally, using simultaneous stimulation of the left and right halves of a face with two different frequencies, we have provided the first objective measure of the integration of parts into a whole face (Boremanse et al., 2013; 2014).
In terms of scientific impact, this work has been presented at many international conferences as well as at an organized symposium on EEG frequency tagging at the International Conference for Cognitive Neuroscience (ICON). The PI has also written a full review papers in on the novel approach (Rossion, 2014; Norcia et al., 2015), and has devoted a section on this approach in a recent general TICS review on the electrophysiology of face perception. The research in other laboratories has been directly inspired by this novel approach and started to use it, leading to independent publications from our laboratory Finally, the Sweep VEP paradigm for high-level stimuli has been developed thanks to our collaboration with A.M. Norcia (Stanford University), by presenting at a periodic the stimuli at a periodic frequency while parametrically varying ("sweeping") the visibility of the stimulus. Thanks to this approach, we were able to determine objective thresholds for face detection for the first time (Ales et al., 2012; Liu-Shuang et al., 2015a; 2015b).

Reported by

Follow us on: RSS Facebook Twitter YouTube Managed by the EU Publications Office Top