Animals and humans deprived of vision have been shown to have enhanced nonvisual perceptual abilities. Indeed, many blind individuals are extremely efficient in tactile processing, including Braille reading; the talent manifested by some well-known blind musicians, singers, and even piano tuners is often in part attributed to the fact that they were blind since their youth. One may conclude that blind persons should be better in nonvisual tasks since they compensate for their lack of vision by focusing on their remaining modalities. Many studies have in fact shown that some early-blind human subjects outperform sighted persons in nonvisual tasks, such as speech perception [1
], unfamiliar voice recognition [4
], verbal memory [5
], and musical abilities [8
]. Of particular relevance to the present study are data suggesting that some blind individuals show better auditory spatial discrimination [11
] or localization of sound sources than sighted subjects [2
]; however, other studies have failed to show this advantage [14
], raising the question of what may underlie individual differences. In general, the nature of any behavioral enhancement, its extent, and its neural bases are still matters of considerable debate.
Animal studies provide some insight as to the neural substrates underlying such enhanced capacities (reviewed in [16
]). For instance, in cats that had been visually deprived for several years by eyelid suture shortly after birth, the auditory cortical representation expanded into visual areas [17
], and auditory spatial tuning was sharpened in the auditory cortex [18
]. Similarly, in neonatally enucleated rats, electrophysiological recordings showed somatosensory responses in the visual cortex [19
], and the somatosensory cortex showed an enlargement of receptive fields of the cells in some barrels together with an increase of angular sensitivity for deflection in another barrel [20
]. Thus, those experiments indicate a recruitment of the visual cortex for nonvisual tasks, but do not conclusively prove that the enhanced perceptions of the blind rely on the visual cortex.
Several studies using neuroimaging techniques have also established that posterior visual areas in blind individuals may be active during the performance of nonvisual tasks such as Braille reading [21
], memory retrieval [7
], and auditory localization [23
] as well as other auditory functions [25
]. It remains to be established whether recruitment of visual cortices necessarily reflects functional reorganization, or whether it indicates a nonspecific or even pathological response. Indeed, despite numerous studies showing activation in visual areas during nonvisual tasks, the functional significance of this phenomenon has been questioned by some investigators who suggest that the occipital cortex might be nonspecifically coactivated [30
]. If the visual cortex participates in nonvisual functions in the blind, then its activity level should be related to individual differences in behavior, and in effect predict behavioral outcome.
Localization of sound, a very important function for the blind, is one domain in which it is particularly useful to study the cross-modal interactions that may occur following visual deprivation. This task entails integration of binaural and monaural cues to derive spatial information. In accordance with the idea that nonvisual processing can be enhanced in the blind, a prior study demonstrated that a subset of early-blind subjects was more accurate than sighted controls (SIG) at localizing sound sources, specifically when using monaural cues [12
]. These findings provide a clear opportunity to study the nature of visual cortical recruitment, and the extent to which it relates to behavioral improvements.
Thus, in the present study, subjects were first studied in an anechoic chamber using binaural and monaural sound localization tasks. Depending on their performance at the monaural task, they were divided into three groups: (i) early-blind participants who could localize the sounds more accurately than control subjects (early blind with superior performance [EBSP]); (ii) early-blind participants who were unable to localize the sounds any more accurately than controls (early blind with normal performance [EBNP]); and (iii) SIG. The same localization task was next adapted so that it could be carried out within the positron emission tomography (PET) apparatus, using a speaker array which permitted pseudo-free-field presentations [31
]. Two control conditions, for monaural sound localization (MSL) and binaural sound localization (BSL), were used to control for the auditory input and motor responses.
The hypothesis tested was that blind persons showing supranormal performance do so because they recruit visual cortical areas to carry out the task. We therefore predicted that they would show activation in visual areas specifically during the MSL task, and not during the binaural or control tasks. The other blind group, which does not have enhanced MSL ability, should not show this activation pattern. We further hypothesized that the degree of visual cortical activity would be predictive of individual differences in the behavioral performance of the monaural task.