In natural vision, information is acquired actively, by directing gaze toward (fixating on) points of interest. Humans and other primates typically scan a visual scene with a large number of brief fixations, at a rate of 3-5 Hz, separated by rapid ``saccadic'' eye movements. At each fixation, a volley of retinal outputs courses into the system and produces a spatiotemporal pattern of brain activation determined by the interaction of stimulus qualities with properties of neurons within each brain area. Despite these well known dynamics, the difficulties in analyzing visual activity in the context of eye movements has led to the traditional approach of examining responses to stimuli that are presented with the eye held stable, either by systemic paralysis in anesthetized nonhuman subjects, or when studying awake subjects (monkeys or humans), by requirements to fixate steadily during stimulus presentation. We and others have investigated active vision in monkeys by tracking the eyes, and by using the onset of the saccade or the fixation as event triggers for response averaging and/or single trial analysis. These studies clearly show that such active vision exploits non-retinal efference copy signals generated by the oculomotor system, which reset the phase of ongoing excitability fluctuations (oscillations) throughout the visual pathways. Because this phase reset aligns the high excitability phase of ambient neuronal oscillations with the inflow in visual information, we hypothesize that 1) the same reset signal enhances oscillatory phase coherence between neuronal ensembles in different brain regions and that this coherence enhances information flow through the system and augments the brain's representation of objects targeted by eye movements, and 2) the saccade-fixation cycle provides an explicit context for information encoding and transformation, organized by a basic sampling rate (i.e., the rate of eye movements), and incorporating/ organizing both bottom-up hierarchical information processing and top-down when/what/where predictions. Our broad goal is to develop a workable approach to high-resolution study of Active Vision dynamics in humans. We have two specific aims: 1 - Define the impact of Active Saccadic Sampling on the dynamics of visual processing. 2 - Investigate information encoding and transformation during Active Sampling of naturalistic scenes. Effective resolution of the dynamics of active vision is of fundamental importance because, aside from a few other forms, such as pursuit eye movements, the brief, snapshot-like fixation provides the major means of sampling the visual environment. An improved understanding of vision in these terms will have numerous implications for improved mechanistic understanding of visual dysfunction, as well as cognitive deficits that stem from abnormal visual search behavior in a wide range of disorders including Huntington's, Parkinson's, ADHD, Fetal Alcohol Syndrome, Schizophrenia.
Active saccadic sampling of natural scenes is essential to normal vision, and its disruption contributes significantly to a range of cognitive deficits in a range of disorders including Huntington's and Parkinson's diseases, Fetal Alcohol Syndrome, autism and schizophrenia. We propose to develop a new paradigm for studying vision. It will yield information critical to improving the understanding and treatment of these disorders.
| Auksztulewicz, Ryszard; Schwiedrzik, Caspar M; Thesen, Thomas et al. (2018) Not All Predictions Are Equal: ""What"" and ""When"" Predictions Modulate Activity in Auditory Cortex through Different Mechanisms. J Neurosci 38:8680-8693 |
| Ten Oever, Sanne; Schroeder, Charles E; Poeppel, David et al. (2017) Low-Frequency Cortical Oscillations Entrain to Subthreshold Rhythmic Auditory Stimuli. J Neurosci 37:4903-4912 |
| Golan, Tal; Davidesco, Ido; Meshulam, Meir et al. (2016) Human intracranial recordings link suppressed transients rather than 'filling-in' to perceptual continuity across blinks. Elife 5: |
