Fred H. Hamker

About the influence of post-saccadic mechanisms for visual stability on peri-saccadic compression of object location

Peri-saccadic perception experiments have revealed a multitude of mislocalization phenomena. For instance, a briefly flashed stimulus is perceived closer to the saccade target, whereas a displacement of the saccade target goes usually unnoticeable. This latter saccadic suppression of displacement has been explained by a built-in characteristic of the perceptual system: the assumption that during a saccade, the environment remains stable. We explored whether the mislocalization of a briefly flashed stimulus toward the saccade target also grounds in the built-in assumption of a stable environment. If the mislocalization of a peri-saccadically flashed stimulus originates from a post-saccadic alignment process, an additional location marker at the position of the upcoming flash should counteract compression. Alternatively, compression might be the result of peri-saccadic attentional phenomena. In this case, mislocalization should occur even if the position of the flashed stimulus is marked. When subjects were asked about their perceived location, they mislocalized the stimulus toward the saccade target, even though they were fully aware of the correct stimulus location. Thus, our results suggest that the uncertainty about the location of a flashed stimulus is not inherently relevant for compression.

A dynamic model of how feature cues guide spatial attention

We will describe a computational model of attention which explains the guidance of spatial attention by feedback within a distributed network. We hypothesize that feedback within the ventral pathway transfers the target template from prefrontal areas into intermediate areas like V4. The oculomotor circuit consisting of FEF, LIP and superior colliculus picks up this distributed activity and provides a continuous spatial reentry signal from premotor cells. In order to test this hypothesis, we simulate two experiments that require a response given a color cue. The first experiment indicates a parallel feature-based enhancement prior to any spatial selection. If two targets are behaviorally relevant, as in the second experiment, experimental findings indicate that subjects split their attention between two locations containing the searched feature. Our simulation results suggest that the split in attention between two foci is a transient effect occurring during competition. We predict that the time after cue presentation determines the state of this competition and ultimately the distribution of attention at different locations. In addition we provide simulation results to explain how reentrant processing through the oculomotor circuit might lead to variations of the time for target detection in visual search.

Life-long learning cell structures —continuously learning without catastrophic interference

As an extension of on-line learning, life-long learning challenges a system which is exposed to patterns from a changing environment during its entire lifespan. An autonomous system should not only integrate new knowledge on-line into its memory, but also preserve the knowledge learned by previous interactions. Thus, life-long learning implies the fundamental Stability-Plasticity Dilemma, which addresses the problem of learning new patterns without forgetting old prototype patterns. We propose an extension to the known Cell Structures, growing Radial Basis Function-like networks, that enables them to learn their number of nodes needed to solve a current task and to dynamically adapt the learning rate of each node separately. As shown in several simulations, the resulting Life-long Learning Cell Structures posses the major characteristics needed to cope with the Stability-Plasticity Dilemma.

The peri-saccadic perception of objects and space.

Eye movements affect object localization and object recognition. Around saccade onset, briefly flashed stimuli appear compressed towards the saccade target, receptive fields dynamically change position, and the recognition of objects near the saccade target is improved. These effects have been attributed to different mechanisms. We provide a unifying account of peri-saccadic perception explaining all three phenomena by a quantitative computational approach simulating cortical cell responses on the population level. Contrary to the common view of spatial attention as a spotlight, our model suggests that oculomotor feedback alters the receptive field structure in multiple visual areas at an intermediate level of the cortical hierarchy to dynamically recruit cells for processing a relevant part of the visual field. The compression of visual space occurs at the expense of this locally enhanced processing capacity.

The emergence of attention by population-based inference and its role in distributed processing and cognitive control of vision

Technologies such as video surveillance and vision guided robotics require flexible vision systems that interpret the scene according to the current task at hand. Attention has been suggested to play an important role in the process of scene understanding by prioritizing relevant information. However, the underlying processes that allow cognition to guide vision have not been fully explored. Our procedure has its origin in current findings of research in attention. We suggest an approach in which high-level cognitive processes are top-down directed and modulate stimulus signals such that vision is a constructive process in time. Prior knowledge is combined with the observation taken from the image by a population-based inference in order to dynamically update the conspicuity of each feature. Any decision, such as object detection, is based on these distributed conspicuities. We demonstrate this concept on a goal-directed object detection task in natural scenes.

The reentry hypothesis: linking eye movements to visual perception

Cortical organization of vision appears to be divided into perception and action. Models of vision have generally assumed that eye movements serve to select a scene for perception, so action and perception are sequential processes. We suggest a less distinct separation. According to our model, occulomotor areas responsible for planning an eye movement, such as the frontal eye field, influence perception prior to the eye movement. The activity reflecting the planning of an eye movement reenters the ventral pathway and sensitizes all cells within the movement field so the planned action determines perception. We demonstrate the performance of the computational model in a visual search task that demands an eye movement toward a target.

Comparing neural networks: a benchmark on growing neural gas, growing cell structures, and fuzzy ARTMAP

This article compares the performance of some recently developed incremental neural networks with the wellknown multilayer perceptron (MLP) on real-world data. The incremental networks are fuzzy ARTMAP (FAM), growing neural gas (GNG) and growing cell structures (GCS). The real-world datasets consist of four different datasets posing different challenges to the networks in terms of complexity of decision boundaries, overlapping between classes, and size of the datasets. The performance of the networks on the datasets is reported with respect to measure classification error, number of training epochs, and sensitivity toward variation of parameters. Statistical evaluations are applied to examine the significance of the results. The overall performance ranks in the following descending order: GNG, GCS, MLP, FAM.

Computational models of basal-ganglia pathway functions: focus on functional neuroanatomy

Over the past 15 years, computational models have had a considerable impact on basal-ganglia research. Most of these models implement multiple distinct basal-ganglia pathways and assume them to fulfill different functions. As there is now a multitude of different models, it has become complex to keep track of their various, sometimes just marginally different assumptions on pathway functions. Moreover, it has become a challenge to oversee to what extent individual assumptions are corroborated or challenged by empirical data. Focusing on computational, but also considering non-computational models, we review influential concepts of pathway functions and show to what extent they are compatible with or contradict each other. Moreover, we outline how empirical evidence favors or challenges specific model assumptions and propose experiments that allow testing assumptions against each other.

2006 Special Issue: V4 receptive field dynamics as predicted by a systems-level model of visual attention using feedback from the frontal eye field

Visual attention is generally considered to facilitate the processing of the attended stimulus. Its mechanisms, however, are still under debate. We have developed a systems-level model of visual attention which predicts that attentive effects emerge by the interactions between different brain areas. Recent physiological studies have provided evidence that attention also alters the receptive field structure. For example, V4 receptive fields typically shrink and shift towards the saccade target around saccade onset. We show that receptive field dynamics are inherently predicted by the mechanism of feedback in our model. According to the model an oculomotor feedback signal from an area involved in the competition for the saccade target location, e.g. the frontal eye field, enhances the gain of V4 cells. V4 receptive field dynamics can be observed after pooling the gain modulated responses to obtain a certain degree of spatial invariance. The time course of the receptive field dynamics in the model resemble those obtained from macaque V4.

Computational models of spatial updating in peri-saccadic perception

Perceptual phenomena that occur around the time of a saccade, such as peri-saccadic mislocalization or saccadic suppression of displacement, have often been linked to mechanisms of spatial stability. These phenomena are usually regarded as errors in processes of trans-saccadic spatial transformations and they provide important tools to study these processes. However, a true understanding of the underlying brain processes that participate in the preparation for a saccade and in the transfer of information across it requires a closer, more quantitative approach that links different perceptual phenomena with each other and with the functional requirements of ensuring spatial stability. We review a number of computational models of peri-saccadic spatial perception that provide steps in that direction. Although most models are concerned with only specific phenomena, some generalization and interconnection between them can be obtained from a comparison. Our analysis shows how different perceptual effects can coherently be brought together and linked back to neuronal mechanisms on the way to explaining vision across saccades.