Neural network modelling of the influence of channelopathies on reflex visual attention

A cortical model with multi-layers to study visual attentional modulation of neurons at the synaptic level

Visual attention is a selective process of visual information and improves perceptual performance by modulating activities of neurons in the visual system. It has been reported that attention increased firing rates of neurons, reduced their response variability and improved reliability of coding relevant stimuli. Recent neurophysiological studies demonstrated that attention also enhanced the synaptic efficacy between neurons mediated through NMDA and AMPA receptors. Majority of computational models of attention usually are based on firing rates, which cannot explain attentional modulations observed at the synaptic level. To understand mechanisms of attentional modulations at the synaptic level, we proposed a neural network consisting of three layers, corresponding to three different brain regions. Each layer has excitatory and inhibitory neurons. Each neuron was modeled by the Hodgkin-Huxley model. The connections between neurons were through excitatory AMPA and NMDA receptors, as well as inhibitory GABAA receptors. Since the binding process of neurotransmitters with receptors is stochastic in the synapse, it is hypothesized that attention could reduce the variation of the stochastic binding process and increase the fraction of bound receptors in the model. We investigated how attention modulated neurons' responses at the synaptic level on the basis of this hypothesis. Simulated results demonstrated that attention increased firing rates of neurons and reduced their response variability. The attention-induced effects were stronger in higher regions compared to those in lower regions, and stronger for inhibitory neurons than for excitatory neurons. In addition, AMPA receptor antagonist (CNQX) impaired attention-induced modulations on neurons' responses, while NMDA receptor antagonist (APV) did not. These results suggest that attention may modulate neuronal activity at the synaptic level.

Simulation of retinal ganglion cell response using fast independent component analysis

Advances in neurobiology suggest that neuronal response of the primary visual cortex to natural stimuli may be attributed to sparse approximation of images, encoding stimuli to activate specific neurons although the underlying mechanisms are still unclear. The responses of retinal ganglion cells (RGCs) to natural and random checkerboard stimuli were simulated using fast independent component analysis. The neuronal response to stimuli was measured using kurtosis and Treves-Rolls sparseness, and the kurtosis, lifetime and population sparseness were analyzed. RGCs exhibited significant lifetime sparseness in response to natural stimuli and random checkerboard stimuli. About 65 and 72% of RGCs do not fire all the time in response to natural and random checkerboard stimuli, respectively. Both kurtosis of single neurons and lifetime response of single neurons values were larger in the case of natural than in random checkerboard stimuli. The population of RGCs fire much less in response to random checkerboard stimuli than natural stimuli. However, kurtosis of population sparseness and population response of the entire neurons were larger with natural than random checkerboard stimuli. RGCs fire more sparsely in response to natural stimuli. Individual neurons fire at a low rate, while the occasional "burst" of neuronal population transmits information efficiently.

EEG classification of driver mental states by deep learning

Driver fatigue is attracting more and more attention, as it is the main cause of traffic accidents, which bring great harm to society and families. This paper proposes to use deep convolutional neural networks, and deep residual learning, to predict the mental states of drivers from electroencephalography (EEG) signals. Accordingly we have developed two mental state classification models called EEG-Conv and EEG-Conv-R. Tested on intra- and inter-subject, our results show that both models outperform the traditional LSTM- and SVM-based classifiers. Our major findings include (1) Both EEG-Conv and EEG-Conv-R yield very good classification performance for mental state prediction; (2) EEG-Conv-R is more suitable for inter-subject mental state prediction; (3) EEG-Conv-R converges more quickly than EEG-Conv. In summary, our proposed classifiers have better predictive power and are promising for application in practical brain-computer interaction .

Vibrational resonance in a randomly connected neural network

A randomly connected network is constructed with similar characteristics (e.g., the ratio of excitatory and inhibitory neurons, the connection probability between neurons, and the axonal conduction delays) as that in the mammalian neocortex and the effects of high-frequency electrical field on the response of the network to a subthreshold low-frequency electrical field are studied in detail. It is found that both the amplitude and frequency of the high-frequency electrical field can modulate the response of the network to the low-frequency electric field. Moreover, vibrational resonance (VR) phenomenon induced by the two types of electrical fields can also be influenced by the network parameters, such as the neuron population, the connection probability between neurons and the synaptic strength. It is interesting that VR is found to be related with the ratio of excitatory neurons that are under high-frequency electrical stimuli. In summary, it is suggested that the interaction of excitatory and inhibitory currents is also an important factor that can influence the performance of VR in neural networks.

The informational entropy endowed in cortical oscillations

A two-dimensional shadow may encompass more information than its corresponding three-dimensional object. Indeed, if we rotate the object, we achieve a pool of observed shadows from different angulations, gradients, shapes and variable length contours that make it possible for us to increase our available information. Starting from this simple observation, we show how informational entropies might turn out to be useful in the evaluation of scale-free dynamics in the brain. Indeed, brain activity exhibits a scale-free distribution that leads to the variations in the power law exponent typical of different functional neurophysiological states. Here we show that modifications in scaling slope are associated with variations in Rényi entropy, a generalization of Shannon informational entropy. From a three-dimensional object's perspective, by changing its orientation (standing for the cortical scale-free exponent), we detect different two-dimensional shadows from different perception angles (standing for Rényi entropy in different brain areas). We show how, starting from known values of Rényi entropy (easily detectable in brain fMRIs or EEG traces), it is feasible to calculate the scaling slope in a given moment and in a given brain area. Because changes in scale-free cortical dynamics modify brain activity, this issue points towards novel approaches to mind reading and description of the forces required for transcranial stimulation.

Estimation of effective connectivity using multi-layer perceptron artificial neural network

Studies on interactions between brain regions estimate effective connectivity, (usually) based on the causality inferences made on the basis of temporal precedence. In this study, the causal relationship is modeled by a multi-layer perceptron feed-forward artificial neural network, because of the ANN's ability to generate appropriate input-output mapping and to learn from training examples without the need of detailed knowledge of the underlying system. At any time instant, the past samples of data are placed in the network input, and the subsequent values are predicted at its output. To estimate the strength of interactions, the measure of "Causality coefficient" is defined based on the network structure, the connecting weights and the parameters of hidden layer activation function. Simulation analysis demonstrates that the method, called "CREANN" (Causal Relationship Estimation by Artificial Neural Network), can estimate time-invariant and time-varying effective connectivity in terms of MVAR coefficients. The method shows robustness with respect to noise level of data. Furthermore, the estimations are not significantly influenced by the model order (considered time-lag), and the different initial conditions (initial random weights and parameters of the network). CREANN is also applied to EEG data collected during a memory recognition task. The results implicate that it can show changes in the information flow between brain regions, involving in the episodic memory retrieval process. These convincing results emphasize that CREANN can be used as an appropriate method to estimate the causal relationship among brain signals.