Simon Jonathan Thorpe

• D.Phil (Oxon)
• Research Director at Université Toulouse III - Paul Sabatier

Here, we provide an analysis of the microsaccades that occurred during continuous visual search and targeting of small faces that we pasted either into cluttered background photos or into a simple gray background. Subjects continuously used their eyes to target singular 3-degree upright or inverted faces in changing scenes. As soon as the participa...

Background Recordings with tetrodes have proven to be more effective in isolating single neuron spiking activity than with single microwires. However, tetrodes have never been used in humans. We report on the characteristics, safety, compatibility with clinical intracranial recordings in epileptic patients, and performance, of a new type of hybrid...

In the context of word learning, it is commonly assumed that repetition is required for young children to form and maintain in memory an association between a novel word and its corresponding object. For instance, at 2 years of age, children are able to disambiguate word-related situations in one shot but are not able to further retain this newly a...

Here, we provide an analysis of the microsaccades that occurred during continuous visual search and targeting of small faces that we pasted either into cluttered background photos or into a simple gray background. Subjects continuously used their eyes to target singular 3-degree upright or inverted faces in changing scenes. As soon as the participa...

The commonly accepted “simple-to-complex” model of visual processing in the brain posits that visual tasks on complex objects such as faces are based on representations in high-level visual areas. Yet, recent experimental data showing the visual system’s ability to localize faces in natural images within 100ms (Crouzet et al., 2010) challenge the p...

Accurate stimulus onset timing is critical to almost all behavioral research. Auditory, visual, or manual response time stimulus onsets are typically sent through wires to various machines that record data such as: eye gaze positions, electroencephalography, stereo electroencephalography, and electrocorticography. These stimulus onsets are collated...

Unlike familiarity, recollection involves the ability to reconstruct mentally previous events that results in a strong sense of reliving. According to the reinstatement hypothesis, this specific feature emerges from the reactivation of cortical patterns involved during information exposure. Over time, the retrieval of specific details becomes more...

Behavioral studies in humans indicate that peripheral vision can do object recognition to some extent. Moreover, recent studies have shown that some information from brain regions retinotopic to visual periphery is somehow fed back to regions retinotopic to the fovea and disrupting this feedback impairs object recognition in human. However, it is u...

There is currently a replication crisis in many fields of neuroscience and psychology, with some estimates claiming up to 64% of research in psychological science is not reproducible. Three common culprits which have been suspected to cause the failure to replicate such studies are small sample sizes, "hypothesizing after the results are known," an...

Human observers readily detect targets and repetitions in streams of rapidly presented visual stimuli. It seems intuitive that regularly spaced repeating items should be easier to detect than irregularly spaced ones, since regularity adds predictability and in addition has ecological relevance. Here, we show that this is not necessarily the case, a...

The primate visual system has inspired the development of deep artificial neural networks, which have revolutionized the computer vision domain. Yet these networks are much less energy-efficient than their biological counterparts, and they are typically trained with backpropagation, which is extremely data-hungry. To address these limitations, we u...

Human observers readily detect targets in stimuli presented briefly and in rapid succession. Here, we show that even without predefined targets, humans can spot repetitions in streams of thousands of images. We presented sequences of natural images reoccurring a number of times interleaved with either one or two distractors, and we asked participan...

Objective: The mechanisms underlying epileptogenicity in tuberous sclerosis complex (TSC) are poorly understood. Methods: We analysed neuronal spiking activity (84 neurons), fast ripples (FRs), local field potentials and intracranial electroencephalogram during interictal epileptiform discharges (IEDs) in the tuber and perituber of a patient usi...

The primate visual system has inspired the development of deep artificial neural networks, which have revolutionized the computer vision domain. Yet these networks are much less energy-efficient than their biological counterparts, and they are typically trained with backpropagation, which is extremely data-hungry. To address these limitations, we u...

While several studies have shown human subjects’ impressive ability to detect faces in individual images in paced settings (Crouzet et al., 2010), we here report the details of an eye movement dataset in which subjects rapidly and continuously targeted single faces embedded in different scenes at rates approaching six face targets each second (incl...

A number of studies have shown human subjects' impressive ability to detect faces in individual images, with saccade reaction times starting as fast as 100 ms after stimulus onset. Here, we report evidence that humans can rapidly and continuously saccade towards single faces embedded in different scenes at rates approaching 6 faces/scenes each seco...

We present a novel strategy for unsupervised feature learning in image applications inspired by the Spike-Timing-Dependent-Plasticity (STDP) biological learning rule. We show equivalence between rank order coding Leaky-Integrate-and-Fire neurons and ReLU artificial neurons when applied to non-temporal data. We apply this to images using rank-order...

Previous studies have shown that spike-timing-dependent plasticity (STDP) can be used in spiking neural networks (SNN) to extract visual features of low or intermediate complexity in an unsupervised manner. These studies, however, used relatively shallow architectures, and only one layer was trainable. Another line of research has demonstrated -- u...

In 2006 Mitchell demonstrated that implicit memory was robust to decay. He showed that the ability to identify fragments of pictures seen 17 years before was significantly higher than for new stimuli. Is this true only for implicit memory? In this study, we tested whether explicit memory was still possible for drawings (n = 144) that had been prese...

Although it has been demonstrated that visual and auditory stimuli can be recalled decades after the initial exposure, previous studies have generally not ruled out the possibility that the material may have been seen or heard during the intervening period. Evidence shows that reactivations of a long-term memory trace play a role in its update and...

Verifying that a face is from a target person (e.g. finding someone in the crowd) is a critical ability of the human face processing system. Yet how fast this can be performed is unknown. The ‘entry-level shift due to expertise’ hypothesis suggests that - since humans are face experts - processing faces should be as fast – or even faster – at the i...

Recent research has demonstrated that humans are able to implicitly encode and retain repeating patterns in meaningless auditory noise. Our study aimed at testing the robustness of long-term implicit recognition memory for these learned patterns. Participants performed a cyclic/non-cyclic discrimination task, during which they were presented with e...

Correlations between sleep quality, sound imagery and discrimination rates of CNs (measured as a'). Clockwise from the top-left: (A) Correlation between sound imagery (measured using the French version of Willander and Baraldi, 2010) to learning and testing performance. (B) Correlation between self-reported sleep quality (measured from a subset of...

Progression of hit rates for looped and scrambled CNs across the testing session. During testing, looped CNs were presented during half of the blocks, i.e., 5 blocks, and the other half included scrambled CNs. As for learned target CN (Figure 4B), discrimination rate for looped and scrambled trials are above chance from the first block.

Performance during the training session, representing number of runs to reach criterion before moving onto the next stage. Participants who did not reach criterion within 5 runs of any stage discontinued the experiment.

A method of performing unsupervised detection of repeating patterns in a series (TS) of events (E21, E12, E5 ...), comprising the steps of: a) Providing a plurality of neurons (NR1 - NRP), each neuron being representative of W event types; b) Acquiring an input packet (IV) comprising N successive events of the series; c) Attributing to at least som...

A digital electronic circuit (SNN) implementing a spiking neural network comprising: an input unit, (IU) for receiving a series of digital signals (ES) representing respective events and for generating a data packet (PK) representative of N contiguous signals of the series, with 1≤N≤M; - a memory (NM) storing data defining a plurality of neurons, c...

Previous studies have shown that spike-timing-dependent plasticity (STDP) can be used in spiking neural networks (SNN) to extract visual features of low or intermediate complexity in an unsupervised manner. These studies, however, used relatively shallow architectures, and only one layer was trainable. Another line of research has demonstrated - us...

Deep convolutional neural nets (DCNNs) are the most successful computational models of primate ventral visual pathway for object recognition. Interestingly, DCNNs tolerate object variations as humans do [1], use IT-like object representations [2], and match the spatio-temporal dynamics of the ventral pathway [3]. Nevertheless, DCNNs are not biologi...

Neuroimaging studies have shown that faces exhibit a central visual field bias, as compared to buildings and scenes. With a saccadic choice task, Crouzet, Kirchner, and Thorpe (2010) demonstrated a speed advantage for the detection of faces with stimuli located 8° from fixation. We used the same paradigm to examine whether the face advantage, relat...

One of the most amazing features of our brain is its capacity to retain sensory memories for years or even decades. For example, people may recognize the names or faces of classmates fifty years after they have left school (Bahrick et al., 1975) or the title of TV programs fifteen years after their broadcast (Squire et al. 1975). In such cases, it...

Humans can initiate ultrafast saccades towards face as early as 100ms post-stimulus onset (Crouzet & Thorpe, 2010, J Vis). However,other object classes, such as cars or animals, have slower mean saccadic reaction times and lower overall saccadic accuracy in comparison to faces. Based on the theory of Spike Timing Dependent Plasticity, which predict...

Our visual systems undoubtedly process far more than we can actually remember. But what determines whether something is remembered or not? Here, I propose that the key to storage could be the duration of neuronal activation. Even when presented in an RSVP stream at 72 images per second, briefly presented stimuli can be fully processed and activate...

Since its inception, the contextual cueing (CC) paradigm has generated considerable interest in various fields of cognitive sciences because it constitutes an elegant approach to understanding how statistical learning (SL) mechanisms can detect contextual regularities during a visual search. In this article we review and discuss five aspects of CC:...

This chapter describes how spike-based learning may allow future models to integrate the efficiency of the latest computer vision systems with human scene processing, focusing on how spiking neural networks may provide a better understanding of how the brain makes sense of the visual world. It suggests that neurons in the visual system can learn ab...

In this issue of Neuron, Quian Quiroga etal. (2014) show that neurons in the human medial temporal lobe (MTL) follow subjects' perceptual states rather than the features of the visual input. Patients with MTL damage however have intact perceptual abilities but suffer instead from extreme forgetfulness. Thus, the reported MTL neurons could create ne...

Earlier studies suggested that the visual system processes information at the basic level (e.g., dog) faster than at the subordinate (e.g., Dalmatian) or superordinate (e.g., animals) levels. However, the advantage of the basic category over the superordinate category in object recognition has been challenged recently, and the hierarchical nature o...

It was recently claimed that humans can perform above chance at guessing where on a screen an image will appear, even when the location is determined by a hardware random number generator (RNG) after the subject has made their response (Bem, 2011, J Pers Soc Psychol, 100, 407). If true, such results would have serious implications for out own work...

Humans can initiate ultrafast saccades towards faces as early as 100ms post-stimulus onset (Crouzet & Thorpe, 2010, J Vis), and even the mean saccadic reaction time can be as short as 120-130 ms, imposing very serious temporal constraints on the underlying mechanisms. To try and understand which brain structures are involved in triggering these ver...

Sounds such as the voice or musical instruments can be recognized on the basis of timbre alone. Here, sound recognition was investigated with severely reduced timbre cues. Short snippets of naturally recorded sounds were extracted from a large corpus. Listeners were asked to report a target category (e.g., sung voices) among other sounds (e.g., mus...

Spiking neural networks are naturally asynchronous and use pulses to carry information. In this paper, we consider implementing such networks on a digital chip. We used an event-based simulator and we started from a previously established simulation, which emulates an analog spiking neural network, that can extract complex and overlapping, temporal...

In a recent study, we demonstrated that a simple two-layer network of spiking neurons equipped with a novel Spike-Time Dependent Plasticity rule was capable of spontaneously learning to detect complex dynamic visual stimuli (Bichler et al, 2012, Neural Networks, 32, 330-48). The network received spike-like events from a Dynamic Vision Sensor chip t...

Human listeners seem to be remarkably able to recognise acoustic sound sources based on timbre cues. Here we describe a psychophysical paradigm to estimate the time it takes to recognise a set of complex sounds differing only in timbre cues: both in terms of the minimum duration of the sounds and the inferred neural processing time. Listeners had t...

The processes underlying object recognition are fundamental for the understanding of visual perception. Humans can recognize many objects rapidly even in complex scenes, a task that still presents major challenges for computer vision systems. A common experimental demonstration of this ability is the rapid animal detection protocol, where human par...

Individual results (accuracy and median RT) of participants in the object task, while objects were presented in congruent or incongruent context. (DOC)

Individual results (accuracy and median RT) of participants in the two tasks (two target classes per task). (DOC)

Characteristics of the animals and vehicles in the natural images. Every animal and vehicle was manually delineated on all natural images used in the experiment. (A) Geometric centroids of the object(s) obtained for each image of the 2 object categories (animals in red, vehicle in gray). When an image contained multiple objects, the average of thei...

Navigating complex routes and finding objects of interest are challenging tasks for the visually impaired. The project NAVIG (Navigation Assisted by artificial VIsion and GNSS) is directed toward increasing personal autonomy via a virtual augmented reality system. The system integrates an adapted geographic information system with different classes...

Over the past 15 years, we have developed software image processing systems that attempt to reproduce the sorts of spike-based processing strategies used in biological vision. The basic idea is that sophisticated visual processing can be achieved with a single wave of spikes by using the relative timing of spikes in different neurons as an efficien...

Human listeners seem to have an impressive ability to recognize a wide variety of natural sounds. However, there is surprisingly little quantitative evidence to characterize this fundamental ability. Here the speed and accuracy of musical-sound recognition were measured psychophysically with a rich but acoustically balanced stimulus set. The set co...

Navigation, especially in unknown areas, remains a major problem for the visually impaired (VI). Over the past 50 years, a number of electronic travel aids (ETAs) have been developed with the aim of improving the mobility of the VI. Despite the efforts, these systems are rarely used. Although the explanation is likely to be incomplete, it is possib...

A saccadic choice task (Kirchner & Thorpe, 2006) was used to measure word processing speed in peripheral vision. To do so, word targets were accompanied by distractor stimuli, which were random strings of consonants presented in the contralateral visual field. Participants were also tested with the animal stimuli of Kirchner and Thorpe's original s...

A biologically inspired approach to learning temporally correlated patterns from a spiking silicon retina is presented. Spikes are generated from the retina in response to relative changes in illumination at the pixel level and transmitted to a feed-forward spiking neural network. Neurons become sensitive to patterns of pixels with correlated activ...

Orientation and mobility are tremendous problems for Blind people. Assistive technologies based on Global Positioning System (GPS) could provide them with a remarkable autonomy. Unfortunately, GPS accuracy, Geographical Information System (GIS) data and map-matching techniques are adapted to vehicle navigation only, and fail in assisting pedestrian...

Recent experimental work has demonstrated the existence of extremely rapid saccades toward faces in natural scenes that can be initiated only 100 ms after image onset (Crouzet et al., 2010). These ultra-rapid saccades constitute a major challenge to current models of processing in the visual system because they do not seem to leave enough time for...

Certain key stimuli such as animals or human faces can trigger very fast saccades from as early as 100 or 120 ms after image onset (Crouzet et al., J Vis, 2010). Can such fast saccades be seen to other more conventional visual stimuli? Here we directly compared the speed and accuracy of saccades to targets defined by either color or orientation. We...

According to Kirchner and Thorpe (2006), it takes less than 120 ms to detect the presence of an animal in peripheral vision. Here we investigated whether the same applies to a very different kind of visual object: printed words. Using their forced-choice saccade paradigm, we measured the time to initiate a saccade to target stimuli presented in per...

How are we able to recognize visual stimuli that we have not seen for decades? This ability poses a clear challenge to current models of how information is stored in neural circuits. I propose that such long-term memories may depend on the existence of highly selective cortical neurons (effectively grandmother cells) that are so selective that they...

Background / Purpose: We present a novel biologically inspired approach to the generation of selective neuronal responses, capable of extracting complex and overlapping temporally correlated features directly from spike-based dynamic vision sensors.Using a purposebuilt spiking neuron simulation system (XNet), we have simulated the development of...

Welcome to APGV 2011! We are pleased to present you the proceedings of the eighth annual Symposium on Applied Perception in Graphics and Visualization, held in Toulouse, France, August 27-28, 2011. The symposium aims at, on the one hand, using insights from perception to advance the field of visual, auditory and multi-sensory representation while,...

In this paper, we present a novel approach to extract complex and overlapping temporally correlated features directly from spike-based dynamic vision sensors. A spiking neural network capable of performing multilayer unsuper­ vised learning through Spike- Timing-Dependent Plasticity is introduced. It shows exceptional performances at detecting cars...

Humans can make very rapid saccades to faces in natural scenes, with latencies as short as 100–110 ms when the choice is between a target on the left or right (Crouzet et al, 2010 Journal of Vision 10), and only 10–15 ms longer when the target can appear anywhere on the screen. What sort of brain circuits could be involved in triggering these ultra...

Humans can make very rapid saccades to faces in natural scenes, with latencies as short as 100–110 ms when the choice is between a target on the left or right (Crouzet et al, 2010 Journal of Vision 10), and only 10–15 ms longer when the target can appear anywhere on the screen. What sort of brain circuits could be involved in triggering these ultra...

Previous studies have shown that subjects can make extremely rapid saccades toward face targets. When the face can be in just one of two positions left or right of fixation, reliable saccades towards face targets can be initiated in as little as 100–110 ms (Crouzet et al, 2010 Journal of Vision) and even when the face can appear in 16 different loc...

It is generally accepted that a typical visual stimulus will be represented by the activity of many millions of neurons distributed across many regions of the visual cortex. However, there is still a long-running debate about the extent to which information about individual objects and events can be read out from the responses of individual neurons...

Numerous studies have shown quantitative and qualitative differences between foveal and peripheral vision but few studies have investigated object recognition beyond 20í. With natural scenes as stimuli Thorpe et al. (1999) reported that human subjects were able to detect the presence of an animal with a performance of 70% correct at an eccentricity...

In this paper I propose a very radical reform of the taxation system, in which a single flat rate financial transaction tax (FTT) is used to replace the vast majority of existing taxes (including VAT, income tax, taxes on profits...). Existing economic data indicates that a flat rate FTT of 1% would generate far more revenue that is currently gener...

When two images are simultaneously flashed left and right of fixation, subjects can initiate saccades to the side where a face is present in as little as 100-110 ms (Crouzet, Kirchner & Thorpe, submitted). In the present study, we tested how performance is affected by reducing the size of the target region within the image. Six different scales wer...

Since the influential studies by Rosch and colleagues in the 70s, it is generally agreed that the visual system can access information at the basic level (e.g., dogs) faster than at the subordinate (e.g., Chihuahua) or superordinate levels (e.g., animals). However, the advantage of the basic category over the superordinate category in object recogn...

Our ability to detect animals in briefly flashed natural scenes is reduced when a high contrast dynamic mask is presented within about 40 ms and masking is complete at the shortest delays (Bacon-Macé et al.,Vision Research, 2005). It is generally assumed that such masking involves high-level interference. To investigate more precisely when and wher...

This paper focuses on feedforward spiking neuron models of the visual cortex. Essentially, we show that a combination of a temporal coding scheme where the most strongly activated neurons fire first with Spike Timing-Dependent Plasticity leads to a situation where neurons will gradually become selective to visual patterns that are both salient, and...

As human listeners, it seems that we should be experts in processing vocal sounds. Here we present new behavioral data that confirm and quantify a voice-processing advantage in a range of natural sound recognition tasks. The experiments focus on time: the reaction-time for recognition, and the shortest sound segment required for recognition. Our be...

The ventral stream of the primate's visual system, involved in object recognition, is mostly hierarchically organised. Along the hierarchy (from V1, to V2, V4, and IT) the complexity of the preferred stimulus of the neurons increases, while, at the same time, responses are more and more invariant to shift, scale, and finally viewpoint. Several feed...

Ultra-rapid categorization studies analyze human responses to briefly flashed, static natural scenes. Recently, we reported that reaction times can be extremely fast if subjects are asked to move their eyes to the side where an animal had appeared. Accuracy was remarkably good with the fastest reliable saccades occurring in only 120 ms after stimul...

Before a natural sound can be recognized, an auditory signature of its source must be learned through experience. Here we used random waveforms to probe the formation of new memories for arbitrary complex sounds. A behavioral measure was designed, based on the detection of repetitions embedded in noises up to 4 s long. Unbeknownst to listeners, som...

This paper discusses some new suggestions for designing hardware vision systems that take inspiration from spike-based biological image processing. The key idea is to modify already existing Address Event Representation (AER) designs so that there is a periodic reset signal that can be generated every time some predefined proportion of "neurons" ha...

Previous work has demonstrated that the human visual system can detect animals in complex natural scenes very efficiently and rapidly. In particular, using a saccadic choice task, H. Kirchner and S. J. Thorpe (2006) found that when two images are simultaneously flashed in the left and right visual fields, saccades toward the side with an animal can...

A prerequisite for the recognition of natural sounds is that an auditory signature of the sources should be learnt through experience. Here, we used random waveforms to investigate whether implicit learning occurs with repeated exposure. Listeners were asked to discriminate between 1-s samples of running noise and two seamlessly repeated 0.5-s samp...

Finding ones way to an unknown destination, navigating complex routes, finding desired inanimate objects; these are all tasks that can be challenging for the visually impaired. The project NAVIG (Navigation Assisted by artificial VIsion and GNSS) is directed towards increasing the autonomy of visually impaired users in known and unknown environment...