1. Uncovering the Neural Mechanisms Underlying Visual Perception
A central line of the research in our lab. is to decipher the neuronal mechanisms mediating visual perception and higher visual functions. We focus on studying the relation between neural activity and behavioral performance.
Previous lab. studies focused on:
V1 encoding of figure-ground segregation and the relation to the animal's perceptual report (Gilad et al. 2013)
V1 encoding of two versus one objects and the relation to the animal's perceptual report (Gilad et al. 2015)
How face vs. non-face can be discriminated in V1 activity? We studied the population responses and their relation to the animal's behavioral report on face images (Ayzenshtat et al. 2012)
Current projects focus on:
Backward masking: Visual backward masking (BM) is a powerful paradigm where the visibility of a brief stimulus is dramatically reduced when followed by a second stimulus, the mask. The stimulus visibility depends on the time interval between stimulus and mask onset (stimulus-to-mask onset asynchrony; SOA), when SOA becomes shorter, stimulus visibility decreases. To date, the neuronal mechanisms underlying BM in the visual cortex are not well understood.
Multi figure-ground segregation: The neural mechanisms involved in figure-ground segregation show that the V1 neural activity in the figure area is enhanced where the neural activity in the background (BG) is suppressed. However, most of these studies involved just a single object (i.e. figure). In a realistic world, the visual scene is not comprised of a single object embedded within a uniform BG, but rather there are several objects embedded in a noisy BG. Visual recognition of multiple objects can emerge from two different behavioral phenomena: counting or subitizing. Subitizing, unlike counting, is a fast operation that applies to up to four or five objects, in which a rapid perception of the number of objects occurs. While there are many behavioral studies on subitizing, the visual neural mechanisms underlying this phenomenon are unknown. While V1 is involved in the FG-modulation of a single object, however, its role in the task of multiple object segregation - remains unknown.
2. Active vision: the influence of eye movements on vision and neural activity
Vision is not a passive process where light hits the retina and information is subsequently processed. Instead, it involves active scanning and exploration of the environment obtained by eye movements. Our eyes constantly move, focusing on different regions and objects, adjusting to varying levels of light, and even predicting stimulus information in the interplay between the periphery and fovea. Understanding active vision is crucial for deciphering the neural code of visual processing and visual perception
Previous lab. studies (Meitovithz et al. 2011) showed that the stimulus-evoked activity is shifted within the retinotopic map in V1 following the stimulus shift over the retina due to fixational saccades known as microsaccades. Additional lab. papers in relation to ey movements: Gilad et al. 2014; Gilad et al. 2017; Slovin H. 2019.
Movie information: the response to a small luminance dot, following a microsaccade. The VSD response is shifted over the retinotopic map of V1 due to the small microsaccade (single trial). Movie created by Tomer Buhnik.
3. Decoding and reconsrtuction of visual content from brain activity
Another important goal of our team is to decode and reconstruct visual stimuli from brain activity of the visual cortex. This is done using brain inspired modeling. The image below demonstrated the reconstruction of a visual stimulus (pixel-by-pixel) from VSD brain activity in V1 (Zurawel et al. 2016; Macknik et al. 2019).
4. Electrical microstimulation combined with VSDI for artificial vision
To generate artificial vision we investigate the effects of intra-cortical microstimulation (ICMS) in V1 and the propagation to higher visual areas. In a recent publication by Oz et al. (2022) we report on these effects and the relation to ICMS evoked saccades.