A Bayesian method for reducing bias in neural representational similarity analysis

1 In neuroscience, the similarity matrix of neural activity patterns in response to different sensory stimuli or under different cognitive states reflects the structure of neural representational space. Existing methods derive point estimations of neural activity patterns from noisy neural imaging data, and the similarity is calculated from these point estimations. We show that this approach translates structured noise from estimated patterns into spurious bias structure in the resulting similarity matrix, which is especially severe when signal-to-noise ratio is low and experimental conditions cannot be fully randomized in a cognitive task. We propose an alternative Bayesian framework for computing representational similarity in which we treat the covariance structure of neural activity patterns as a hyperparameter in a generative model of the neural data, and directly estimate this covariance structure from imaging data while marginalizing over the unknown activity patterns. Converting the estimated covariance structure into a correlation matrix offers an unbiased estimate of neural representational similarity. Our method can also simultaneously estimate a signal-to-noise map that informs where the learned representational structure is supported more strongly, and the learned covariance matrix can be used as a structured prior to constrain Bayesian estimation of neural activity patterns. 1 Neural pattern similarity as a way to understand neural representations Understanding how patterns of neural activity relate to internal representations of the environment is one of the central themes of both system neuroscience and human neural imaging [20, 5, 7, 15]. One can record neural responses (e.g. by functional magnetic resonance imaging; fMRI) while participants observe sensory stimuli, and in parallel, build different computational models to mimic the brain’s encoding of these stimuli. Neural activity pattern corresponding to each feature of an encoding model can then be estimated from the imaging data. Such activity patterns can be used to decode the content of different brain areas to the encoding process, and to decode perceived stimuli This is not the final version to appear in the proceedings of the 29th NIPS conference (2016). Please cite the most up-to-date reference from NIPS proceedings if possible, or the corresponding journal paper, once published. 29th Conference on Neural Information Processing Systems (NIPS 2016), Barcelona, Spain. . CC-BY-NC-ND 4.0 International license peer-reviewed) is the author/funder. It is made available under a The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/073932 doi: bioRxiv preprint first posted online Sep. 7, 2016;