Mgr. Ján Antolík, Ph.D.

Mgr. Ján Antolík, Ph.D.

Ing. Magda Friedjungová, Ph.D.

Department of Applied Mathematics

This thesis explores the application of deep learning techniques for reconstructing visual stimuli from neural activity in the primary visual cortex (V1). The focus is placed on overcoming the scarcity of biological data by developing data-efficient architectures, analyzing the impact of synthetic training data, employing adversarial and transfer learning, and introducing novel auxiliary optimization objectives. A series of experiments is conducted using data from in silico simulations of cat V1 and in vivo recordings from mouse V1, highlighting the best-performing decoding approach and offering suggestions for future research. Notably, the methods developed in this thesis outperform some existing state-of-the-art decoding techniques according to several widely used evaluation measures. Overall, the results underscore the potential of machine learning in neural activity decoding and pave the way for future advancements in brain-computer interfaces and neuroscientific research.

Thesis on DSpace

Mgr. Ján Antolík, Ph.D.

Ing. Jakub Šístek, Ph.D.

Department of Applied Mathematics

This thesis extends a multi-stage recurrent neural network of the primary visual cortex with a connection learning module. The original model learns weights for all lateral connections, which yields a quadratic number of free parameters. We have given the model functional properties of all neurons and a small DNN, which assigns weights based on these properties. Our solution reduces the number of free parameters from O(n^2) to O(n). We also show that the learned weights are high for neurons with similar properties like orientation or phase, which is consistent with the biological structure of the visual cortex.

Thesis on DSpace