Cosyne 2009 Workshops
March 2-3, 2009
Snow Bird, Utah
Workshop Title
Prefrontal contributions to visual perception
Organizer(s)
Hakwan Lau, Columbia University
Abstract
Formal analysis has shown that the prefrontal cortex is anatomically the final converging point of the dorsal and ventral stream of visual processing. However, most empirical investigation of visual perception is focused on posterior areas of the brain. This bias is recently beginning to be rectified, partly due to new data collected with whole-brain methods such as fMRI. Electrophysiologists are also beginning to record neural activity from prefrontal areas. These new findings call for new ideas regarding what role the prefrontal cortex plays in vision. Some have incorporated existing computational ideas into their empirical research. Some are developing new models to incorporate these new data.
The aim of the workshop is to bring together people conducting these lines of research, to compare differences in their findings and approaches, and to draw general principles underlying the different aspects of prefrontal involvement in visual perception. We also intend to raise awareness of this new focus of research, because it opens up new opportunities for the conceptualization and modelling of higher cognitive aspects of vision. The new prefrontal cortex findings allow us to think about some of the most interesting issues of visual perception, such as the dynamics of attentional modulation, sensorimotor transformation, uncertainty monitoring, perceptual decision making, predictive coding, etc. It is an exciting new area for theoretical research. The target audience is therefore both computational modelers who are interested in these higher cognitive aspects of perception, as well as vision scientists who are relatively familiar with the traditional posterior visual processes, but wish to extend their focus to the prefrontal cortex in order the understand the visual brain as a whole. People who study other prefrontal functions such as working memory and cognitive control should also find the workshop interesting.
Speakers
8:30am-8:35am Opening remarks
8:35am-9:10am Hakwan Lau (Columbia University; organizer)
Title: The role of the prefrontal cortex in making confidence or subjective visibility ratings
Abstract: Humans, monkeys, as well as non-primate subjects can give subjective reports such as confidence or visibility ratings when making perceptual decisions. We developed psychophysical paradigms under which the subjective ratings can be dissociated from the objectively measured capacity for perceptual processing (e.g. d' in signal detection theoretic terms) - subjects claimed that they saw more or were more certain of their decisions in some conditions even when the performance was the same. Using fMRI, we showed that when perceptual capacity was matched, activity in the dorsolateral prefrontal cortex reflects the subjective ratings. Transcranial magnetic stimulation (TMS) to this regions impaired subject's ability to report their ratings properly. Formal comparison of computational models suggests that subjective ratings depend on a late stage of information processing in a hierarchy (instead of depending on a separate channel, or the same information that drives the perceptual decision). We argue that the prefrontal cortex supports processing at this late stage.
9:10am-9:20am Questions and Discussion
9:20am-9:55am Christopher Summerfield (University of Oxford)
Title: A priori visual information in the orbitofrontal cortex
An observer deciding whether a signal is present or absent must account for the probable costs or benefits associated with 'yes' and 'no' responses. But how does the estimated value of each alternative bias perceptual choice? Here, building competing 'drift-diffusion' models of signal detection, we demonstrate that rewarding 'yes' over 'no' responses (engendering a 'liberal' bias) leads subjects to increase their a priori estimates of the probability of signal presence. An a priori model of biased perceptual choice captured real human agents' tendency to make fast, impulsive choices under liberal bias conditions. Moreover, during functional magnetic resonance imaging (fMRI), a priori estimates of signal presence generated by this model predicted brain activity in the orbitofrontal cortex, a region implicated in the representation of economic value, and top-down control over visual perception.
9:55am - 10:05am Questions and discussion
10:05am - 10:15am Coffee break
10:15am - 10:50am Basilis Zikopoulos (Boston University)
Title: Circuits for multisensory integration and attentional modulation through the prefrontal cortex and the thalamic reticular nucleus
Abstract: How do we focus attention on a task in a complex environment with multiple distractions? Acting as a central executive, the prefrontal cortex selects relevant information and discards irrelevant information. We recently reported that specific areas of the prefrontal cortex send widespread pathways to the thalamic reticular nucleus, which intercepts and filters all communications between the thalamus and the cortex. Prefrontal pathways impinge on parts of the reticular nucleus that receive information from sensory areas, suggesting that they can influence the flow of information from the senses passing from the thalamus to the cortex. Unlike sensory pathways, the prefrontal pathways form some large synapses in the reticular nucleus, suggesting efficient transfer of information. The widespread prefrontal pathways in the reticular nucleus may help focus attention, select relevant information and override distractors, functions that are deranged in schizophrenia and attention deficit disorder.
10:50am - 11am Questions and discussion
4:30pm - 5:05pm Tatiana Pasternak (U of Rochester)
Title: What does prefrontal cortex "know" about visual motion used in discrimination tasks?
Abstract: Perceptual decisions during visual discrimination tasks often require subjects to compare two or more sequentially presented stimuli. During such tasks the stimuli not only have to be processed, but also retained in memory and the comparison between the remembered and the current stimulus has to be performed. To characterize the cortical circuitry sub-serving successful execution of such tasks we focused on speed and direction of visual motion and compared the behavior of neurons during such tasks in two interconnected cortical regions, the motion processing area MT and a region associated with executive control and working memory, prefrontal cortex (PFC). During these tasks, the monkeys compared either directions or speeds of two moving random-dot stimuli, sample and test, separated by a brief memory delay. During direction discrimination, many PFC neurons showed robust direction selectivity, reminiscent of MT neurons. The same neurons were tuned for stimulus speed but reduced their DS when the monkeys discriminated stimulus speeds and ignored its direction. This loss in DS, most pronounced in putative inhibitory inter-neurons, point to a mechanism capable of enhancing selectivity for behaviorally relevant sensory signals. During the memory delay, both MT and PFC neurons carried reliable motion signals, although in individual neurons such signals were largely transient and decreased with time, suggesting that the contribution of these neurons to stimulus maintenance could only be accomplished at the population level. During the test, responses in both areas were modulated by the preceding direction, reflecting access to the remembered sample. In PFC, these comparison-related signals were in the form of match suppression and appeared 100ms later than similar signals in MT, pointing to MT as a likely source of sensory comparison. Thus, during motion discrimination tasks, neurons in areas MT and PFC make unique contributions to different task components and are likely to be functionally linked, a hypothesis supported by recent recordings from MT deprived of "top-down" influences from the ipsilateral PFC.
5:05pm-5:15pm Questions and discussion
5:15pm-5:50pm Kirk Thompson (Neuroscience Section, National Institute of Health)
Title: The role of the frontal eye field in visual perception
Abstract: The frontal eye field (FEF) lies at the interface between visual and oculomotor processing. There is growing evidence that FEF not only converts visual information into commands to move the eyes, but also actively participates in visual processing. I will present results of recent neurophysiological experiments in behaving monkeys exploring the role of FEF in object identification without eye movements. Our results suggest that spatially selective activity in FEF is necessary for monkeys to accurately report the identity of visual objects during a visual search task among distractors, but not when the objects are presented alone. These results indicate an important role for spatial vision in general, and specifically for FEF in the processing of visual objects along the ventral visual processing stream.
5:50pm-6:00pm Questions and Discussion
6:00pm-6:15pm Coffee break
6:15pm-6:50pm Vincent Ferrara (Columbia University)
Title: Frontal Eye Fields: Integration of form and motion for saccade planning and decision-making.
Abstract: The frontal eye field (FEF) receives inputs from both the dorsal and ventral visual pathways (Schall et al. 1995). Recently, FEF neurons have been found to have robust shape selectivity (Peng et al. 2008), as well as selectivity for direction and speed of motion (Xiao et al. 2006). They can also exhibit selectivity for features such as color when they are linked to specific motor responses (Bichot et al. 1996; Ferrera et al. 1999). These observations raise questions such as 1) how are form and motion information integrated in for purposes of saccade planning, and 2) can FEF play a role in functions that are thought to be specific to the domain of object vision, such as categorizing a visual stimulus independently of a specific motor response? We trained monkeys to perform various eye movement tasks and then recorded neuronal activity in their FEF. In one task, a small moving target was “deflected” by an oriented bar and the monkey had to make an eye movement to the predicted future location of the target. This location depended not only on the motion of the target, but also on the orientation of the bar, neither cue alone was sufficient to successfully perform the task. Monkeys learned to accurately predict the location of the moving target. Neuronal activity in FEF showed selectivity well in advance of the eye movement. In some cases, the tuning emerged before the target hit the deflector. In a second experiment, we asked whether FEF was capable of playing a role in categorical decision-making. To investigate this, we developed a speed categorization task in which monkeys were presented with a random dot motion stimulus. They were required to make a saccadic eye movement to one of two response targets to indicate whether the stimulus was “fast” or “slow”. The task was designed so that monkeys associated the speed categories with the colors of the response targets. The locations of the response targets were randomized. Hence, the categorical decision was independent of the motor response. The monkeys were able to learn two different category boundaries and to shift rapidly between them. Activity in FEF during stimulus presentation was significantly modulated both by stimulus speed and by boundary speed in 34 and 44% of the neurons, respectively. Some cells encoded both the physical speed of the stimulus and its category, which could be considered a categorical representation. This pattern of activity allowed a simple rule to be used to read out neural activity: neurons were considered as casting a vote for their preferred speed category if their firing rate was higher than average, and a vote for their non-preferred category if their firing rate was lower than average. This rule could be used to categorize stimuli on a trial-by-trial basis. This method of reading out neuronal activity closely approximated the monkeys’ behavioral performance. The results of the two experiments suggest that FEF activity reflects the integration of form and motion cues, and also that FEF can perform functions such as categorization that are thought to be in the domain of the ventral visual pathway.
6:50pm-7pm Questions and Discussion
