Associative learning in early vision

Learning by Exposure in the Visual System

It is increasingly being understood that perceptual learning involves different types of plasticity. Thus, whereas the practice-based improvement in the ability to perform specific tasks is believed to rely on top-down plasticity, the capacity of sensory systems to passively adapt to the stimuli they are exposed to is believed to rely on bottom-up plasticity. However, top-down and bottom-up plasticity have never been investigated concurrently, and hence their relationship is not well understood. To examine whether passive exposure influences perceptual performance, we asked subjects to test their orientation discrimination performance around and orthogonal to the exposed orientation axes, at an exposed and an unexposed location while oriented sine-wave gratings were presented in a fixed position. Here we report that repetitive passive exposure to oriented sequences that are not linked to a specific task induces a persistent, bottom-up form of learning that is stronger than top-down practice learning and generalizes across complex stimulus dimensions. Importantly, orientation-specific exposure learning led to a robust improvement in the discrimination of complex stimuli (shapes and natural scenes). Our results indicate that long-term sensory adaptation by passive exposure should be viewed as a form of perceptual learning that is complementary to practice learning in that it reduces constraints on speed and generalization.

Stimulus roving and flankers affect perceptual learning of contrast discrimination in Macaca mulatta

'Stimulus roving' refers to a paradigm in which the properties of the stimuli to be discriminated vary from trial to trial, rather than being kept constant throughout a block of trials. Rhesus monkeys have previously been shown to improve their contrast discrimination performance on a non-roving task, in which they had to report the contrast of a test stimulus relative to that of a fixed-contrast sample stimulus. Human psychophysics studies indicate that roving stimuli yield little or no perceptual learning. Here, we investigate how stimulus roving influences perceptual learning in macaque monkeys and how the addition of flankers alters performance under roving conditions. Animals were initially trained on a contrast discrimination task under non-roving conditions until their performance levels stabilized. The introduction of roving contrast conditions resulted in a pronounced drop in performance, which suggested that subjects initially failed to heed the sample contrast and performed the task using an internal memory reference. With training, significant improvements occurred, demonstrating that learning is possible under roving conditions. To investigate the notion of flanker-induced perceptual learning, flanker stimuli (30% fixed-contrast iso-oriented collinear gratings) were presented jointly with central (roving) stimuli. Presentation of flanker stimuli yielded substantial performance improvements in one subject, but deteriorations in the other. Finally, after the removal of flankers, performance levels returned to their pre-flanker state in both subjects, indicating that the flanker-induced changes were contingent upon the continued presentation of flankers.

Visual perception of ambiguous figures: synchronization based neural models

We develop and study two neural network models of perceptual alternations. Both models have a star-like architecture of connections with a central element connected to a set of peripheral elements. A particular perception is simulated in terms of partial synchronization between the central element and some sub-group of peripheral elements. The first model is constructed from phase oscillators and the mechanism of perceptual alternations is based on chaotic intermittency under fixed parameter values. Similar to experimental evidence, the distribution of times between perceptual alternations is represented by the gamma distribution. The second model is built of spiking neurons of the Hodgkin-Huxley type. The mechanism of perceptual alternations is based on plasticity of inhibitory synapses which increases the inhibition from the central unit to the neural assembly representing the current percept. As a result another perception is formed. Simulations show that the second model is in good agreement with behavioural data on switching times between percepts of ambiguous figures and with experimental results on binocular rivalry of two and four percepts.

Attention alters visual plasticity during exposure-based learning

It is generally believed that attention enhances the processing of sensory information during perception and learning. Here we report that, contrary to common belief, attention limits the degree of plasticity induced by repeated exposure to image features. Specifically, daily exposure to oriented stimuli that are not linked to a specific task causes an orientation-specific improvement in perceptual performance along the "exposed" axes. This effect is modulated by attention: human subjects showed a larger improvement in orientation discrimination when attention is directed toward the location where stimuli are presented. However, the capacity to perform discriminations away from the exposed orientation is enhanced when the exposure stimuli are unattended. Importantly, the improvement in orientation discrimination at the unattended location leads to a robust enhancement in the discrimination of complex stimuli, such as natural texture images, with orientation components along the exposed axes, whereas the improvement in orientation discrimination at the attended location exhibits only weak transfer to complex stimuli. These results indicate that sensory adaptation by passive stimulus exposure should be viewed as a form of perceptual learning that is complementary to practice-based learning in that it reduces constraints on generalization.

Stimulus coding rules for perceptual learning

Perceptual learning of visual features occurs when multiple stimuli are presented in a fixed sequence (temporal patterning), but not when they are presented in random order (roving). This points to the need for proper stimulus coding in order for learning of multiple stimuli to occur. We examined the stimulus coding rules for learning with multiple stimuli. Our results demonstrate that: (1) stimulus rhythm is necessary for temporal patterning to take effect during practice; (2) learning consolidation is subject to disruption by roving up to 4 h after each practice session; (3) importantly, after completion of temporal-patterned learning, performance is undisrupted by extended roving training; (4) roving is ineffective if each stimulus is presented for five or more consecutive trials; and (5) roving is also ineffective if each stimulus has a distinct identity. We propose that for multi-stimulus learning to occur, the brain needs to conceptually “tag” each stimulus, in order to switch attention to the appropriate perceptual template. Stimulus temporal patterning assists in tagging stimuli and switching attention through its rhythmic stimulus sequence.

When a person learns to judge several stimuli in succession, like baseball pitches arriving at various speeds and spins, judgments may improve with practice only if these stimuli are presented in a fixed temporal sequence, rather than in a random order. These contrary effects suggest the need for proper stimulus coding for multi-stimulus learning in the brain. We studied how the temporal order of the stimuli affects the encoding, consolidation, and retrieval stages of perceptual learning that describe the basic stimulus coding rules throughout the learning process. We also studied why fixed stimulus sequences are required for multi-stimulus learning. Our results suggest that for multi-stimulus learning to occur, the brain needs to identify or tag each stimulus conceptually or semantically, so that the neural activity specific to each stimulus can be properly attended to. This high-level conceptual process adds to the current understanding of the mechanisms underlying perceptual learning and may have important implications for sensory training and rehabilitation.

Visual discrimination skills are more easily learned when stimuli are presented in a fixed rhythm.

Defining a link between perceptual learning and attention

Takeo Watanabe and Yuko Yotsumoto explore the implications of a new study that shows that for perceptual learning of visual features involving multiple stimuli to occur, the brain needs to temporally "tag" the features, a learning process that requires paying attention.

Perceptual learning via modification of cortical top-down signals

The primary visual cortex (V1) is pre-wired to facilitate the extraction of behaviorally important visual features. Collinear edge detectors in V1, for instance, mutually enhance each other to improve the perception of lines against a noisy background. The same pre-wiring that facilitates line extraction, however, is detrimental when subjects have to discriminate the brightness of different line segments. How is it possible to improve in one task by unsupervised practicing, without getting worse in the other task? The classical view of perceptual learning is that practicing modulates the feedforward input stream through synaptic modifications onto or within V1. However, any rewiring of V1 would deteriorate other perceptual abilities different from the trained one. We propose a general neuronal model showing that perceptual learning can modulate top-down input to V1 in a task-specific way while feedforward and lateral pathways remain intact. Consistent with biological data, the model explains how context-dependent brightness discrimination is improved by a top-down recruitment of recurrent inhibition and a top-down induced increase of the neuronal gain within V1. Both the top-down modulation of inhibition and of neuronal gain are suggested to be universal features of cortical microcircuits which enable perceptual learning.

Asymmetric synaptic depression in cortical networks

Synaptic depression is essential for controlling the balance between excitation and inhibition in cortical networks. Several studies have shown that the depression of intracortical synapses is asymmetric, that is, inhibitory synapses depress less than excitatory ones. Whether this asymmetry has any impact on cortical function is unknown. Here we show that the differential depression of intracortical synapses provides a mechanism through which the gain and sensitivity of cortical circuits shifts over time to improve stimulus coding. We examined the functional consequences of asymmetric synaptic depression by modeling recurrent interactions between orientation-selective neurons in primary visual cortex (V1) that adapt to feedforward inputs. We demonstrate analytically that despite the fact that excitatory synapses depress more than inhibitory synapses, excitatory responses are reduced less than inhibitory ones to increase the overall response gain. These changes play an active role in generating selective gain control in visual cortical circuits. Specifically, asymmetric synaptic depression regulates network selectivity by amplifying responses and sensitivity of V1 neurons to infrequent stimuli and attenuating responses and sensitivity to frequent stimuli, as is indeed observed experimentally.

The essential role of stimulus temporal patterning in enabling perceptual learning

Little is known about how temporal stimulus factors influence perceptual learning. Here we demonstrate an essential role of stimulus temporal patterning in enabling perceptual learning by showing that 'unlearnable' contrast and motion-direction discrimination (resulting from random interleaving of stimuli) can be readily learned when stimuli are practiced in a fixed temporal pattern. This temporal patterning does not facilitate learning by reducing stimulus uncertainty; further, learning enabled by temporal patterning can later generalize to randomly presented stimuli.

Learning prosthetic vision: a virtual-reality study

Acceptance of prosthetic vision will be heavily dependent on the ability of recipients to form useful information from such vision. Training strategies to accelerate learning and maximize visual comprehension would need to be designed in the light of the factors affecting human learning under prosthetic vision. Some of these potential factors were examined in a visual acuity study using the Landolt C optotype under virtual-reality simulation of prosthetic vision. Fifteen normally sighted subjects were tested for 10-20 sessions. Potential learning factors were tested at p < 0.05 with regression models. Learning was most evident across-sessions, though 17% of sessions did express significant within-session trends. Learning was highly concentrated toward a critical range of optotype sizes, and subjects were less capable in identifying the closed optotype (a Landolt C with no gap, forming a closed annulus). Training for implant recipients should target these critical sizes and the closed optotype to extend the limit of visual comprehension. Although there was no evidence that image processing affected overall learning, subjects showed varying personal preferences.

Perceptual learning: specificity versus generalization

Perceptual learning improves performance on many tasks, from orientation discrimination to the identification of faces. Although conventional wisdom considered sensory cortices as hard-wired, the specificity of improvement achieved through perceptual learning indicates an involvement of early sensory cortices. These cortices might be more plastic than previously assumed, and both sum-potential and single cell recordings indeed demonstrate plasticity of neuronal responses of these sensory cortices. However, for learning to be optimally useful, it must generalize to other tasks. Further research on perceptual learning should therefore, in my opinion, investigate first, the conditions for generalization of training-induced improvement, second, its use for teaching and rehabilitation, and third, its dependence on pharmacological agents.

Neural networks and perceptual learning

Sensory perception is a learned trait. The brain strategies we use to perceive the world are constantly modified by experience. With practice, we subconsciously become better at identifying familiar objects or distinguishing fine details in our environment. Current theoretical models simulate some properties of perceptual learning, but neglect the underlying cortical circuits. Future neural network models must incorporate the top-down alteration of cortical function by expectation or perceptual tasks. These newly found dynamic processes are challenging earlier views of static and feedforward processing of sensory information.