Thesis Type:PhD thesis
We live in a society based on vision. Visual information is used for orienting in our environment, identifying objects in our surroundings, alerting us to important events which require our attention, engaging in social interactions, and many more functions that are necessary to efficiently function in everyday life. Thus, the loss of vision decreases the quality of life and poses a severe challenge to efficient functioning for millions of individuals worldwide.
Despite some medical progress, the restoration of visual information to the blind still faces multiple technical and scientific difficulties. “Bionic eyes” or visual prostheses are being developed mostly for specific blindness etiologies and target only a subpopulation of the visually impaired. Even these devices have yet to reach the stage where the technology can provide high-resolution, detailed visual information.
More importantly, these approaches take for granted the ability of the human brain, following long-term or even life-long blindness, to interpret vision once the input from the eyes becomes available.
The current scientific consensus regarding the development of the visual cortex is that visual deprivation during critical or sensitive periods in early development may result in functional blindness, as the brain is not organized to process visual information properly, and this may be irreversible later in life. In fact, the rare reported cases of late-onset surgical sight restoration (by means of cataract removal in early blind patients) show severe visual impairments that persist even following long-term exposure to vision. Furthermore, studies of early-onset and congenitally blind people have shown that their visual cortex may have plastically reorganized to process information from other sensory modalities. Recent studies showed that even short term visual deprivation in adulthood may cause some functional chances in the visual system. Thus, sight restoration may indeed be severely limited by the reorganization of the visual cortex.
In this dissertation I test this theory by using an alternative approach to visual rehabilitation, in which the visual information is conveyed non-invasively using the remaining senses of the blind. We used a sensory substitution device (SSD) that translates visual information using a consistent algorithm to sounds (The vOICe). Because SSD soundscape translations of natural visual input are very complex, we developed a structured training protocol (Striem-Amit et al., 2012b) where congenitally blind people are gradually taught how to interpret the sounds carrying the visual information. This training paradigm also made it possible to test whether the congenitally blind can learn to perceive complex visual information without having had visual experience during early infancy, and to better identify the neural correlates of processing such information in the blind.
Specifically, this dissertation aimed to study:
1. Whether and how SSDs may be applied for visual rehabilitation to reach sufficient practical visual acuity and functional abilities. Can we restore complex visual capacities such as object categorization (a visual ability which requires feature binding within visual objects as well as their segregation from their background) beyond the critical developmental period in the congenitally blind?
2. How are such visual-in-nature artificially-constructed stimuli processed in the blind brain? Can we find evidence for the functional specializations of the normal visual cortex in the absence of visual experience during the critical periods of early development?
We found that the blind were able to learn to perceive high-acuity visual information, and could even exceed the Snellen acuity test threshold of the World Health Organization for blindness (Striem-Amit et al., 2012d) and the 'visual' acuity possible using any other current means of visual rehabilitation. Furthermore, they were able to perceive and categorize images of visual categories, and carry out certain visual tasks (Striem-Amit et al., 2012b).
A neuroimaging investigation of the processing of SSD information showed that despite their lack of visual experience during development, the visual cortex of the congenitally blind was activated during the processing of soundscapes (images represented by sounds). More importantly, its activation pattern mimicked the task- and category- selectivities of the normally developed visual cortex. Specifically, we found that the blind showed a double-dissociation between processing image shape and location in the ventral and dorsal processing streams (Striem-Amit et al., 2012c), which constitutes the large-scale organization principle in the visual cortex. Furthermore, we found that within the ventral stream, category-selectivity for one visual category over all others tested can be seen in the visual word-form area (VWFA) which, as in the normally sighted, showed a robust preference for letters over textures and other visual categories (Striem-Amit et al., 2012b).
In both studies, the visual cortex showed retention of functional selectivity despite the a-typical auditory sensory-modality input, the lack of visual experience, the limited training duration (dozens of hours) and the fact that such training was applied only in adulthood. These findings support a controversial organization theory which suggests that instead of being divided according to the sensory modalities which elicit it, the cortex area may be better defined by the tasks or computations it conducts, whereas the input sense organ is irrelevant. Specifically, this model suggests that a combination of top-down connectivity with an innate preference for computation- type may generate the same task-selectivities even in the absence of bottom-up visual input. This theory has interesting bearings on the ability to restore sight later in life, as it suggests that the blind brain may not have lost its ability to process some aspects of visual information and may learn to do so if this information is delivered to the brain either via SSDs, as we show possible here, or using other more invasive means.