Karen Hampson, Imperial College, United Kingdom
Adaptive optics (AO) is becoming an increasingly popular technique to investigate the human visual system such as the possible role of the higher-order aberrations in the accommodation response. Firstly this talk will outline the design considerations and performance of a system constructed at Imperial College. Then results will be presented on three main experiments carried out using the system.
The higher-order aberrations of the human eye are known to fluctuate over relatively short time periods but the origin of the fluctuations has not been determined with certainty. The first set of experiments was to determine if the heartbeat is a possible cause. The aberrations and pressure pulse wave were simultaneously measured for five subjects. Coherence function analysis was used to assess the correlations.
Frequency entrainment is when an oscillation at one frequency in a system causes another oscillation at a close frequency to be pulled until eventually both oscillate at the same frequency. An example of this is brain wave entrainment via a light source stimulus with certain properties. The second set of experiments involved investigating if the fluctuations in the higher-order aberrations and accommodation could be entrained to certain frequencies governed by the mirror.
The final experiments discussed are those in which the possible role of the higher-order aberrations in the accommodation response was investigated. These experiments involved selectively implementing or correcting certain aberrations.
Wei Ji Ma, California Institute of Technology
We will discuss two examples of how the brain can manipulate uncertainty in processing brief stimuli.
- When a 10 ms flash is presented along with two 10 ms beeps, observers often report seeing two flashes. This illusion is a consequence of the way the brain combines sensory information. When a bisensory task requires observers to report a single percept, the content of this percept has been shown to be determined by a maximum-likelihood rule. In the flash-beep paradigm, we asked for both a visual and an auditory report. A Bayesian model can then account for the full range of responses, from complete fusion via cross-modal modulation to complete segregation.
- Visual short-term memory is commonly modeled as a box which can hold about four items reliably, while items which fall outside are subject to guessing. We have conducted a change detection experiment with confidence ratings; this allows to plot receiver-operator characteristics. These do not at all correspond to the predictions of a box model. Instead, a signal detection model can explain the results very well.
Iain Gilchrist, Department of Experimental Psychology, University of Bristol, UK
Saccadic eye movements are a ubiquitous part of visual behaviour. The classical view is that we move our eyes in order to keep the world static on the retina or to point the high acuity fovea to regions of interest (e.g. Walls, 1962). However, because saccadic eye movements sample limited parts of the visual environment over time, they may serve a more fundamental computational function. I will provide support for this idea using data from visual search and a simple reasoning task.
Steven Jax, Graduate student, Pennsylvania State University
Current theories of manual obstacle avoidance do not predict sequential effects in obstacle avoidance because the theories assume, explicitly or implicitly, that the way a movement is planned does not depend on what movements were performed before. The goal of the present research was to determine whether sequential effects do in fact occur in reaching around obstacles. We asked participants to perform reaching movements between targets in the presence of or in the absence of an intervening obstacle. Results from four experiments showed significant sequential effects. When obstacle-present and obstacle-absent trials were intermixed within a block of trials, the hand paths in the obstacle-absent trials were more curved than when no obstacle was ever present. This obstacle-perseveration (OP) effect appears to be due to difficulties in switching from an obstacle-avoidance mode to a non-obstacle-avoidance mode of performance. In support of this interpretation, the effect does depend on the probability of obstacle-present trials (Experiment 2) but does not depend on the randomness of trial switches (Experiments 3 and 4).
Peggy Series, Postdoc, Brain & Cognitive Sciences, University of Rochester
Many neurons in cortex exhibit bell-shaped tuning curves. Several studies have shown that the information conveyed by these tuning curves increases as their width decreases, leading to the notion that sharpening tuning curves improves population codes-- a principle invoked by models in a wide variety of domains, including orientation selectivity, perceptual learning, attention, and auditory processing. This notion, however, is based on the assumptions that the noise distribution (trial-to-trial variability) is 1) independent among neurons, and 2) independent of the tuning curve width. We have reexamined these assumptions in networks of spiking neurons using orientation selectivity as an example. We compared the two major classes of models: a model in which the tuning curves are sharpened through cortical lateral interactions and a model in which the orientation selectivity is the result of the convergence of LGN afferences with no further sharpening in the cortex. We report that the sharpening model conveys far less information about orientation than the no-sharpening model. Therefore, contrary to what is widely assumed, sharpening through lateral connections in networks of spiking neurons leads to severe losses of information. Moreover, the code produced by the sharpening model is particularly inefficient for learning and computation because the majority of the information is conveyed by correlations. This work also makes several experimental predictions that could be used to distinguish between the sharpening and no-sharpening models.
Ted Dryja, Harvard Medical School and Massachusetts Eye and Ear Infirmary
Inherited diseases of the photoreceptor cells of the retina, such as retinitis pigmentosa, cone dystrophy, congenital retinal blindness, and stationary night blindness, are clinically and genetically heterogeneous. Most of these diseases are monogenic and have classical inheritance patterns, such as autosomal dominant, autosomal recessive, X-linked, or mitochondrial (maternally inherited), while others have more complex inheritance patterns or are multifactorial. In the last 14 years, about 80 identified genes have been reported; 39% of these genes are specifically expressed by the retina and the remainder are expressed by the retina as well as by many other tissues. The encoded proteins participate in many biochemical pathways, such as phototransduction, RNA transcription, RNA splicing, and the production or maintenance of cytoskeletal structures. There are over 45 additional genes that remain unidentified, most of which have been assigned to chromosomal regions through linkage studies. The frequency of recessive retinitis pigmentosa and the large number of responsible genes indicates that the aggregate carrier frequency may be over 10%. The identification of two genes that cause a novel photoreceptor phenotype, bradyopsia, illustrates my laboratory's approach to the identification of retinal disease genes.
Marc Schieber, Assoc. Professor, University of Rochester
The operation of a real neural network can be examined by studying the network of connections between the primary motor cortex (M1) and spinal motoneuron pools. Many M1 neurons make connections to multiple spinal motoneuron pools (muscles); and each motoneuron pool receives inputs that converge from many M1 neurons. Which M1 neurons provide inputs to which muscle(s) can be determined with spike-triggered averaging of rectified electromyographic activity. We then can compare the extent to which the activity patterns of M1 neurons, each weighted by the strength of its functional connection to a given muscle, can be summed to account for the activity pattern of that muscle during 12 individuated finger and wrist movements. The M1 neurons that provide inputs to any given muscle show diverse patterns of activity. Some are partially similar to that of the muscle, others are quite dissimilar. The sum of M1 neuron activity patterns, each weighted by the strength of its connection to the muscle, therefore only partially resembles that of the muscle. Shifting the muscle pattern back in time by 40 to 60 msec typically increases the similarity, suggesting that M1 neurons with relatively direct connections to motoneuron pools also influence the same motoneuron pools through more indirect pathways. Residual differences between the summed neuron activity pattern and muscle activity pattern may indicate variation in the effective neuron-muscle connection strength, nonlinear summation of inputs and/or inputs from subcortical centers.
Matt Dye, Postdoc, University of Rochester
Most research on the development of visual attention skills in school-aged children has focused upon filtering, orienting and visual search. This research has suggested that improvement is seen during the school years in a child's ability to suppress task irrelevant information, reorient after invalid spatial cues, and search for targets in a visual scene. In the study reported here we administered four test of visual attention to 79 school children aged 7-18 years. The effects of age and action video-gaming experience on filtering and orienting were examined using a Posner-style task (Attentional Network Test), as were their effects on the spatial distribution of attention Useful Field of View), the time required to recover after expending attentional resources (Attentional Blink), and the ability to simultaneously track multiple objects (Ball Tracking). Results suggest that all aspects of visual attention develop with age. Some -- such as orienting -- are impervious to the effects of action gaming, whereas others skills -- spatial distribution of attention and the ability to track multiple objects -- are clearly enhanced by such an activity. Time required to recover attentional resources was modified by action gaming experience, but only for specific age groups.
Hal Greenwald, Graduate student, University of Rochester
We used a novel psychophysical technique to evaluate the separate contributions of monocular and binocular cues over time to planning and online control of guided reaching movements. This is different from most visual cue integration studies, which have relied on static perceptual judgments and thus have been unable to analyze the dynamics of cue integration. Like other voluntary motor behaviors, reaching movements consist of two phases, planning and online control, and our experiment examined how information about visual slant (orientation in depth) contributes to each. Also, since reaches are relatively fast movements, they enabled us to examine how processing time impacts cue integration. Subjects placed a cylinder flush onto a circular target surface that varied randomly in slant between trials while viewing a binocular image of the surface and cylinder in a 3D virtual environment. Conflicts between monocular cues (contour, texture) and binocular cues (disparity, vergence) were either present in the stimulus at initial presentation, affecting both planning and online control, or added after movement initiation (a "perturbation"), affecting only the online control component of the movement. The results indicated that in the perturbation trials, binocular cues influenced final cylinder orientation more than did monocular cues and, for unperturbed trials, binocular cues remained more influential but to a lesser degree. A temporal analysis of cue weights revealed that the influence of binocular information on subjects' movements accrued faster than that of monocular information, which was consistent with the finding that final cylinder orientation was more strongly correlated with the binocular cue for fast than for slow movements. The results also showed that humans appear to give more weight to binocular cues for online control than for planning. This can be explained by our finding that binocular cues are processed more quickly than monocular cues, effectively giving them more influence over the course of what are relatively short movements.
Bill Geisler, Center for Perceptual Systems and Department of Psychology, University of Texas at Austin
Generic visual search in the real world involves integrating information over multiple fixations, yet most research has focused on single-fixation tasks where stimuli are presented briefly. At least two factors have held back progress in understanding visual processing in more natural search tasks: (1) the difficulty of precisely controlling and manipulating the stimulus on the retina and (2) the lack of an ideal-observer theory for multiple-fixation visual search. To enable precise stimulus control in extended search tasks, we developed gaze-contingent software that allows real-time control of the content of video displays relative to the observer's current gaze direction (measured with an eye tracker). We have used this software to investigate how much information can be removed from the periphery (how much foveation can be tolerated) without affecting search time or the pattern of eye movements. To provide an appropriate benchmark against which to evaluate search performance, and to provide a starting point for developing models of search performance, we derived the ideal observer for visual search in broadband noise, where the ideal searcher is constrained by an arbitrary function describing sensitivity across the retina and by some level of internal noise. We compare the eye movements and performance of ideal and real observers. We find that a number of aspects of human search performance and eye movement patterns are qualitatively consistent with the ideal searcher.
Uri Polat, Tel-Aviv University, Goldschleger Eye Research Institute
The functional architecture and plasticity of lateral interactions in the visual system was explored using psychophysics, visual evoked potentials and single unit recordings. Our results show that interactions depend on stimulus configuration and activity level. At low activity level, long-range excitatory interactions, organized mainly along the neuron's optimal orientation, were found to be most effective. At higher activity levels shorter range and broadband inhibition dominated local activity. Insight on developmental factor affecting lateral interactions was obtained by studying human amblyopia. Amblyopia is characterized by several functional abnormalities in spatial vision, including reduction in visual acuity and contrast sensitivity function, resulting from abnormal development of one eye's cortical representation during the "critical period". The damage was thought to be irreparable after the first decade of life once the developmental maturation window has been terminated. In amblyopia, we found lateral interactions to be imbalanced due to a high degree of inhibition and reduced amount of excitation. A perceptual learning procedure was developed to train this network by efficiently stimulating these neuronal populations and effectively promoting their spatial interactions. The training procedure improved the abnormal spatial interactions and produced a two-fold improvement in contrast sensitivity and visual acuity. The results demonstrate that perceptual learning can improve basic representations within the adult visual system that did not develop during the critical period.
Marla Feller, Professor, Division of Biological Sciences, UC San Diego
At very early stages in visual system development, even before the retina is responsive to light, spontaneous retinal activity plays a vital role in the establishment of adult circuits. Spontaneous activity in the developing retina occurs as waves of action potentials that correlate the firing of neighboring retinal neurons with a periodicity on the order of minutes. I will describe two sets of experiments that address the mechanisms and function of these highly correlated activity patterns. First, to elucidate the cellular mechanisms for slow periodic activity, we cultured dissociated retinal neurons. Whole cell voltage clamp recordings reveal high levels of spontaneous synaptic activity that occur in bursts with an interval of minutes. Preliminary experiments indicate that this slow periodicity is not an emergent property of a complex neural network by rather is driven by cell-autonomous pacemakers. Second, we used transgenic mice with altered patterns of spontaneous activity to dissect what are the salient features of retinal waves for driving the activity-dependent organization of retinothalamic projections.
Elena Fedorovskaya, Eastman Kodak Co.
I will present the framework and results of experiments aiming at developing computational methods to predict perceived image quality and perceived attributes, based on digital image data. Subjective judgments of overall contrast, sharpness, lightness, colorfulness, and quality were collected for 126 images presented on a CRT display. For the analysis, digital images were represented in CIELAB color space, which was used as an approximation of a perceptually uniform color space. We utilized a computational vision approach and showed that the global perceptual attributes, including quality, can be predicted based on the combination of features (image properties) that could be considered as relevant elements at different levels of mental representation. The dependence of the global attributes upon the same features explains their mutual correlations.
David Heeger, Professor, Psychology and Neural Science, New York University
When the two eyes view large dissimilar patterns that induce binocular rivalry, alternating waves of visibility are experienced, as one pattern sweeps the other out of conscious awareness. We have used fMRI to examine neural responses accompanying perceptual waves within early visual cortex. Latencies and amplitudes of fMRI signals in areas V1, V2 and V3 varied dynamically across the retinotopic map and correlated with the latency, direction and quality of perceptual waves. Similar patterns of signal fluctuations were found when composite monocular patterns were shown to mimic perceptual waves that observers saw during rivalry. Within V1, but not in V2 or V3, the existence of neural waves during rivalry was maintained when attention was diverted from the inducing pattern, making it impossible for observers to track the waves perceptually. Our fine-scale analysis of fMRI signals reveals a tight linkage between the dynamics of binocular rivalry and neural events in early visual cortical areas including V1, but activity within higher visual areas seems required for waves to be perceived.
Jason Porter, Institute of Optics, University of Rochester
Laser in-situ keratomileusis (LASIK) procedures correct the eye's defocus and astigmatism but also introduce higher order monochromatic aberrations. Little is known about the origins of these induced aberrations. The advent of wavefront sensor technology has made it possible to measure accurately and quickly the aberrations of normal and postoperative LASIK eyes. The goal of this thesis was to exploit this technology to better understand some of the potential mechanisms by which aberrations could be introduced during LASIK.
A first step towards investigating these sources was to characterize the aberration changes in post-LASIK eyes. Higher order rms wavefront error increased after conventional and customized LASIK surgery. On average, spherical aberration approximately doubled, and significant changes in vertical and horizontal coma were observed.
We examined two sources of postoperative aberrations: the creation of a microkeratome flap and the subsequent laser ablation. Higher order rms increased slightly and there was a wide variation in the response of individual Zernike modes after cutting a flap. The majority of induced spherical aberration was due to the laser ablation and not the flap-cut.
Aberrations are also induced by static and dynamic decentrations of the patient's pupil. We found that ablations were typically decentered in the superotemporal direction due to shifts in pupil center location between aberration measurement (dilated) and surgical (undilated) conditions in customized LASIK eyes. There was a weak correlation between the horizontal coma theoretically induced by this offset and that measured postoperatively.
Finally, dynamic eye movements during the procedure induce higher order aberrations. We found that the most problematic decentrations during LASIK are relatively slow drifts in eye position. An eye-tracking system with a 2-Hz closed-loop bandwidth could compensate for most eye movements during LASIK. One solution for reducing the aberrations induced by static and dynamic shifts in pupil center location is to reference the aberration measurement and treatment with respect to fixed features on the eye. Several other sources of aberration induction in LASIK, such as the efficiency of laser pulses striking the cornea perpendicularly vs. obliquely, must still be investigated to optimize postoperative optical quality after LASIK.
Rebecca Sappington, Neuroscience, University of Rochester
Glaucoma is a common optic neuropathy that is characterized by progressive loss of retinal ganglion cells (RGCs) and generally correlates with chronic increases in intraocular pressure. The anatomical intimacy between RGCs and retinal glia suggests that extracellular signaling may contribute to death of RGCs in glaucoma. Pressure-induced responses of RGCs, astrocytes and microglia are likely to involve both an intrinsic component specific to each cell type and an extrinsic component dependent on signals between cell types. The broad objective of this work is to apply an in vitro model of glaucoma to elucidate the contribution of both intrinsic and extrinsic factors to RGC death. This thesis focuses on pressure-induced apoptosis of RGCs and B and subsequent production of cytokines by retinalkactivation of NF glia as well as modulation of these responses to pressure by ion channels with putative mechanosensitive gating.
We developed an in vitro model of glaucoma where immuno-magnetically purified cultures of primary RGCs, retinal astrocytes and retinal microglia are exposed to elevated hydrostatic pressure. We measured pressure-induced apoptotic death of RGCs aB and production of IL-6 and TNFkand activation of NF by retinal glia. We found that RGCs, astrocytes and microglia each respond to elevated pressure in a cell-type specific manner. RGCs are inherently more susceptible to pressure-induced apoptosis than other neurons. Similarly, elevated pressure that is differentiallyaB activation and production of IL-6 and TNFkinduces NF modulated by astrocytes and microglia anda. Furthermore, pressure-induced release of TNF IL-6 by retinal glia promotes RGC survival. Thus, pressure-induced death of RGCs can be modulated by both intrinsic and extrinsic signals.
The differential response of retinal cells to pressure suggests that each detects and transduces a direct pressure signal. We found that RGCs, astrocytes and microglia express genes for the transient receptor potential (TRP) family of mechanically-gated ion channels. Furthermore, the response of each cell type to pressure was modulated by calcium and inhibition of TRP channels. These data suggest that expression and activation of TRP channels may modulate the response of retinal cells to elevated pressure and that these channels could serve as possible therapeutic targets in glaucoma.
Seth Pantanelli, Biomedical Engineering, University of Rochester
The Shack-Hartmann wavefront sensor has become a powerful tool for diagnostic purposes in vision science. Despite its widespread use, an inherent dynamic range limitation prevents conventional devices from measuring largely aberrated eyes, such as those having keratoconus or penetrating keratoplasty (PK). Keratoconus is a progressive disease that results in a thin, cone-shaped cornea. Penetrating keratoplasty patients are those that have had a corneal transplant. Since these eyes stand to benefit most from developing customized correction methods, it is imperative that robust methods for wavefront measurement and characterization exist for these populations.
A novel large dynamic range Shack-Hartmann wavefront sensor is presented. The modified design, which involves placing a translatable plate in conjugate with the lenslet array, blocks every other spot on the photodetector in a single exposure. All of the spots are collected by translating the plate to reveal previously masked lenslets. The device increases the dynamic range by a factor of two without sacrificing any measurement sensitivity. With this device, the aberration of abnormal eyes can be measured over a large pupil.
The wavefront was measured and Zernike coefficients calculated for 19 keratoconic and 14 PK eyes. The aberration data was averaged for each population. From the analysis, obvious trends and similarities became apparent. All keratoconic subjects except for one possessed negative vertical coma; the magnitude of the aberration was -1.41 ± 1.06 µm (mean ± standard deviation). Penetrating keratoplasty subjects exhibited large amounts of trefoil; however there was no consistency with the sign. Overall, the higher-order variance in both abnormal eye populations accounted for 16% of the total variance, compared to only 1% for normal eyes. Visual benefit calculations for 12 of the keratoconic eyes indicate that they may experience visual acuity improvements ranging between 2 and 10 times what is possible with conventional correction.
A new large dynamic range Shack-Hartmann wavefront sensor has been successfully developed to characterize the aberrations and optical image quality in two groups of abnormal eyes over a large pupil (6.0 mm). By measuring the aberrations in a relatively large population, a more comprehensive understanding of the image quality degradation is obtained. Incorporating this device into the clinic will ultimately provide better diagnostic and therapeutic treatment of ocular pathologies. Additionally, visual performance may be substantially improved with developing customized correction methods, such as customized contact lenses or customized intraocular lenses.
Dennis Dacey, University of Washington, Seattle
Matthew V. Chafee, Neuroscience, University of Minnesota
Parietal neurons in the monkey have long been recognized to have visual receptive or movement fields, regions of space where visual input or motor output is associated with physiological activation. These data have established the role of parietal cortex in sensorimotor integration. More recent work has shown that cognitive operations are also associated with neural activity in primate posterior parietal cortex, specifically with the modulation of the receptive or movement field properties of parietal neurons. In studies where animals attended to stimuli or decided where to move, attention, intention, and decision processing were each linked to changes in either the intensity or the timing of neural signals that coded the location of stimuli or the direction of movements. I will address the question whether parietal neurons participate in abstract spatial cognitive processing, by which I mean the representation of spatial parameters that are computed in order to achieve a cognitive objective, but that do not reflect the locations of stimuli or directions of movement. The data I will present argue that the nature of the cognitive demand placed on parietal cortex has an impact on the nature of the neural representation of space generated to meet that demand.
Douglas Goodman, Sr. Scientist, Corning Tropel
The OSA holds an Educator's Day at its annual meeting during which about one hundred high school science teachers learn things about optics and vision that they can use in the classroom. Doug Goodman, this year's organizer, will be here to describe the program and recruit CVS people who might contribute to this effort. This is an opportunity to advertise CVS/UR and to provide a good mechanism for networking. In addition, it is an excellent way to stimulate lots of young minds about a potential future in vision science. CVS participation will also help to ensure factual accuracy for non-vision experts.
Goodman will show a few demonstrations representative of those to be shown at the event, and he will present the department with an interesting phenomenon in which reflection produces the illusion of a gooey liquid-y substance.
Alireza Soltani, Physics, Brandeis University
Sheila Nirenberg, Neurobiology, UCLA
Our research focuses on the general question "How do networks of neurons carry out computations?", and concentrates largely on the visual system. While a great deal is known about individual neurons in this system - in terms of their response properties and patterns of connections - relatively little is known about how they act together in networks to produce specific outputs or carry out computations. Even computations that seem relatively simple, such as determining the direction of motion of a moving object or the orientation of a line are still not completely understood. One approach to this problem is to dissect networks using targeted cell class ablation - that is, systematically eliminate different classes of cells from a network and assess how its output is perturbed. The focus of my talk will be on a method we developed for this purpose and the results of using it to study basic network operations.
Alex Wade, UCSF
Special RCBI Event
Martin Banks, UC Berkeley School of Optometry
Alexandre Pouget, Brain and Cognitive Sciences, University of Rochester
The spike count of cortical neurons in response to a fixed stimulus follows a near Poisson distribution. This variability is often referred to as Poisson noise, which implies that it is detrimental to computation. We propose instead that it is a computational feature that allows cortical circuits to perform Bayesian inferences through simple linear transformations. This idea predicts that neurons involved in decision making should behave like neural integrator, which is precisely what has been reported in area LIP. I will show how this approach can be applied to a variety of domains such as cue integration, decision making and motor control.
Okihide Hikosaka, Laboratory of Sensorimotor Research, NEI
Saccadic eye movement is controlled by many brain areas including the frontal and parietal cortices, cerebellum, and basal ganglia. All of these areas project to the superior colliculus (SC) to determine the goal of a saccade. Why do we need these parallel pathways? Which of these areas decides to initiate a saccade? To answer these questions, we recorded single neurons in basal ganglia nuclei using behavioral paradigms in which the decision to make a saccade was influenced by cognitive and motivational factors. It has been known that presumed GABAergic projection neurons in the caudate nucleus and the substantia nigra pars reticulata exhibit sensory, cognitive, and saccadic motor responses which would guide a saccade to a position in space. We found that these spatial signals were frequently and strongly modulated by the presence/absence of expected reward and/or the position in space where reward was expected. Presumed dopaminergic neurons in and around the substantia nigra exhibited reward-selective but spatially non-selective sensory responses. These results suggest that spatial signals in GABAergic caudate neurons (which may originate from the frontal and parietal cortices) are modulated or conditioned by reward-related dopaminergic inputs. We speculate that the basal ganglia are preferentially involved in decision making based on motivational values.
I will discuss reverberatory cortical microcircuits with the ability to both maintain working memory and perform decision making computations. Endowed with reward-gated Hebbian plasticity, the same networks can learn arbitrary sensori-motor associations underlying flexible behavior.
Barry Lee, SUNY College of Optometry
We are concerned with the nature and precision of motion signals emanating from retinal ganglion cells. It turns out that the signals of the transient, magnocellular (MC) pathway are spatially accurate even at low drift rates, both for discrete targets and gratings. MC cell behavior matches quite closely psychophysical performance on vernier tasks with moving targets. However, there are some interesting differences, and we attempt to use this information to help constrain central processing mechanisms. Finally, results with grating targets containing both luminance and chromatic contrast throw further light on the nature of spatial mechanisms processing motion.
Krystel Huxlin, University of Rochester
Damage to the adult visual cortex causes largely permanent visual impairment. However, our recent behavioral and anatomical studies in cats demonstrate that recovery of visual motion perception is possible following cortical lesions when intensive but localized visual re-training is administered to the impaired visual field. Motion perception is critical for navigating and interacting with the environment, and is affected by damage at different levels of the visual system. To assess whether improvements in visual motion perception could be elicited in adult humans with primary visual cortical damage, three patients were recruited one year after a stroke that induced homonymous visual field defects. They were taught to self-administer a two-alternative, forced-choice, global direction discrimination task within their blind fields. On a daily basis, they performed 300 trials of this task using random-dot stimuli that are optimally processed in higher-level visual areas such as the MT/MST complex, which was intact in these patients. Slow but significant recovery of global motion perception was observed over 2-3 months. Recovery was highly specific to the visual field location(s) retrained and was paralleled by increasing conscious perception of the visual stimulus. Patients also demonstrated improved ability to detect and track moving objects in a realistic virtual environment. Thus, direction discrimination retraining with dynamic random dot stimuli improves perceptual motion thresholds and the usage of visual motion information in naturalistic environments after permanent visual cortical damage in adulthood.
Christof Koch, California Institute of Technology
The ability to record the simultaneous spiking activity of many neurons in behaving subjects, combined with the large scale coverage possible with functional brain imaging, has enabled neurobiologists to begin to study the neuronal basis of subjective, conscious experience. In this context, Francis Crick and I advocate characterizing the neuronal correlates of consciousness (NCC). I will outline a two-pronged research program to discover the NCC that our laboratory is pioneering: the study of visual consciousness in humans using psychophysics, fMRI and multi-electrode single neuron recordings and the study of contingency awareness in mice using aversive Pavlovian conditioning.
Jamie Hillis, University of Pennsylvania
Color supports two distinct functions of vision. The first is surface discrimination: color variation across the image is an important cue for determining what regions of the image belong to distinct surfaces. The second is identification: color, as the perceptual correlate of surface reflectance, provides an important cue to identity. To preserve information key to accomplishing these two tasks, the visual system adapts to prevailing viewing conditions. However, the demands of each function on adaptation are different: (1) To preserve discriminability, response differences to light reflected from different surfaces must be maintained (2) to preserve identity, response to light from the same surface under different conditions must be fixed. Given this difference, it seems worthwhile to ask: Are color appearance and discriminability controlled by the same mechanisms of adaptation? I will discuss recent theoretical and empirical work aimed at addressing this question.