Purchase this article with an account.
Petra Vetter, Fraser W. Smith, Lars Muckli; Decoding natural sounds in early visual cortex. Journal of Vision 2011;11(11):779. doi: https://doi.org/10.1167/11.11.779.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Despite early visual cortex being one of the most intensely studied systems in the brain, the majority of processing variance is still unexplained. Here we show that information from hearing is contained in the activity patterns of early visual cortex, in a content-specific manner, in the absence of visual stimulation and despite an orthogonal working memory task. In Exp. 1, subject listened to 3 types of natural sounds while being blindfolded. We successfully decoded the sounds from retinotopically defined early visual cortex BOLD activity (mainly in V2 and V3) using multivariate pattern analysis (a linear support vector machine). To constrain mental imagery, subjects performed an orthogonal working memory task in Exp. 2, again in the absence of visual stimulation. Subjects memorised word lists of animals and objects during natural sound stimulation and performed a delayed match to sample task. Natural sounds were still successfully decoded from early visual cortex activity, despite mental imagery being restricted. These results show that contextual information from the auditory modality is contained in the activity patterns of early visual cortex, possibly to facilitate visual predictions.
This PDF is available to Subscribers Only