|
|
Decoding visual stimuli from cortical activity
Vašek, Vojtěch ; Antolík, Ján (advisor) ; Šikudová, Elena (referee)
This thesis aims to develop a machine learning model that can decode stimu- lus images from cortical activity in the primary visual cortex (V1) to understand the relationship between V1 activity and visual perception. The limited avai- lability of biological data makes it necessary to use the spiking neural network model of V1 to generate the underlying training data. Machine learning tech- niques, particularly neural networks, will be explored to generate high-quality stimulus images. Standard loss functions, as well as discriminator loss from GAN networks training, will be used to train the decoding models. Linear regression models will be used baseline. The research questions to be addressed include the best decoding approach, the impact of the number of neurons recorded or stimuli presented, the loss of information in high frequencies domain and the effect of intrinsic noise in neural responses on reconstructing visual stimuli. This thesis proposes a trainable convolutional network, which outperforms other baseline models such as linear regression. We observe that the loss function producing the best results is the MSSSIM. However, the intrinsic noise in neural respon- ses limits the reconstruction, and only low frequencies are being reconstructed. The size of the dataset and the number of cortical...
|
|
|
Coding of pheromone signal by olfactory receptor neurons in Agrotis ipsilon
Kováčová, Kristýna ; Košťál, Lubomír (advisor) ; Pokora, Ondřej (referee)
i Abstract The main objective of the thesis is to describe differences in the activity of male A. ipsilon olfactory receptor neurons (ORNs) when stimulated by different temporal dynamics of the concentration of the conspecific female pheromone. First, under the artificial situation of constant pulse stimulation, and second, with a fluctuating signal resembling the natural situation. For this purpose, the experimental data were collected in the collaborating laboratory (Dr. P. Lucas, INRAe, Versailles, France) by employing a novel olfactometer system that enables precise temporal control of the pheromone delivery to individual sensilla. Using the R programming language, we analyzed various descriptors of the response reliability, randomness, and variability, as well as the information content of the evoked activity. The results are interpreted in the context of the classical efficient coding hypothesis, which states that sensory neurons are evolutionarily adapted to natural stimuli. The main finding is that although the response variability is widely spread across the ORN population, sometimes with no visible difference between the constant and fluctuating stimulation types, the fluctuating stimulus is usually encoded with systematically higher reliability, as revealed by the inspection of individual ORNs....
|
|
|
Special Issue on the 18th International Conference on Artificial Neural Networks
Húsek, Dušan ; Neruda, Roman ; Koutník, J.
Special Issue on the 18th International Conference on Artificial Neural Networks. Neural Network World. Vol. 19, No. 5 (2009). The issue contains papers prepared specially for this issue by authors of some best evaluated papers presented on ICANGA 2008 conference. Covered are mainly following topics: Mathematical Theory of Neurocomputing, Computational Neuroscience, Connectionist Cognitive Science, Neuroinformatics, Image Processing, Signal and Time Series Processing, Reinforcement Learning, Binary Factor Analysis, Principal Component Analysis, Self-organization, Neural Network Hardware.
|
guest ::
