The ability for visual prostheses to preferentially activate functionally-distinct retinal ganglion cells (RGCs) is important for improving visual perception. This study investigates the use of high frequency stimulation (HFS) to elicit RGC activation, using a closed-loop algorithm to search for optimal stimulation parameters for preferential ON and OFF RGC activation, resembling natural physiological neural encoding in response to visual stimuli. We evaluated the performance of a wide range of electrical stimulation amplitudes and frequencies on RGC responses in vitro using murine retinal preparations. It was possible to preferentially excite either ON or OFF RGCs by adjusting amplitudes and frequencies in HFS. ON RGCs can be preferentially activated at relatively higher stimulation amplitudes (>150 μA) and frequencies (2–6.25 kHz) while OFF RGCs are activated by lower stimulation amplitudes (40–90 μA) across all tested frequencies (1–6.25 kHz). These stimuli also showed great promise in eliciting RGC responses that parallel natural RGC encoding: ON RGCs exhibited an increase in spiking activity during electrical stimulation while OFF RGCs exhibited decreased spiking activity, given the same stimulation amplitude. In conjunction with the in vitro studies, in silico simulations indicated that optimal HFS parameters could be rapidly identified in practice, whilst sampling spiking activity of relevant neuronal subtypes. This closed-loop approach represents a step forward in modulating stimulation parameters to achieve appropriate neural encoding in retinal prostheses, advancing control over RGC subtypes activated by electrical stimulation.

Various time-series decomposition techniques, including wavelet transform, singular spectrum analysis, empirical mode decomposition and independent component analysis, have been developed for non-linear dynamic system analysis. In this paper, we describe a symplectic geometry spectrum analysis (SGSA) method to decompose a time series into a set of independent additive components. SGSA is performed in four steps: embedding, symplectic QR decomposition, grouping and diagonal averaging. The obtained components can be used for de-noising, prediction, control and synchronization. We demonstrate the effectiveness of SGSA in reconstructing and predicting two noisy benchmark nonlinear dynamic systems: the Lorenz and Mackey-Glass attractors. Examples of prediction of a decadal average sunspot number time series and a mechanomyographic signal recorded from human skeletal muscle further demonstrate the applicability of the SGSA method in real-life applications.

The vertebrate retina is a clearly organized signal-processing system. It contains more than 60 different types of neurons, arranged in three distinct neural layers. Each cell type is believed to serve unique role(s) in encoding visual information. While we now have a relatively good understanding of the constituent cell types in the retina and some general ideas of their connectivity, with few exceptions, how the retinal circuitry performs computation remains poorly understood. Computational modeling has been commonly used to study the retina from the single cell to the network level. In this article, we begin by reviewing retinal modeling strategies and existing models. We then discuss in detail the significance and limitations of these models, and finally, we provide suggestions for the future development of retinal neural modeling.

Retinal electrostimulation is promising a successful therapy to restore functional vision. However, a narrow stimulating current range exists between retinal neuron excitation and inhibition which may lead to misperformance of visual prostheses. As the conveyance of representation of complex visual scenes may require neighbouring electrodes to be activated simultaneously, electric field summation may contribute to reach this inhibitory threshold. This study used three approaches to assess the implications of relatively high stimulating conditions in visual prostheses: (1) in vivo, using a suprachoroidal prosthesis implanted in a feline model, (2) in vitro through electrostimulation of murine retinal preparations, and (3) in silico by computing the response of a population of retinal ganglion cells. Inhibitory stimulating conditions led to diminished cortical activity in the cat. Stimulus-response relationships showed non-monotonic profiles to increasing stimulating current. This was observed in vitro and in silico as the combined response of groups of neurons (close to the stimulating electrode) being inhibited at certain stimulating amplitudes, whilst other groups (far from the stimulating electrode) being recruited. These findings may explain the halo-like phosphene shapes reported in clinical trials and suggest that simultaneous stimulation in retinal prostheses is limited by the inhibitory threshold of the retinal ganglion cells.

Objective. This study focused on characterising the response of four major functionally-different retinal ganglion cells (RGCs) to a high frequency stimulus (HFS) paradigm. Approach. We used in vitro patch clamp experiments to assess the viability of evoking a differential response between different RGC types—OFF-Sustained, OFF-Transient, ON-Sustained and ON-Transient—under a wide range of HFS and stimulation amplitude combinations. Main results . Of the four types, we found that the OFF-Sustained, OFF-Transient and ON-Transient RGCs could be differentially activated at various frequency and amplitude combinations, in particular, OFF-Sustained cells can be differentially targeted between 20–100 µ A at all frequencies, OFF-Transient cells between 150–240 µ A at 1 kHz and ON-Transient between 180–240 µ A and 4–6 kHz. We further found that this differential activation held true when the stimulus duration was reduced from 300 ms to 50 ms. Finally, we found that the cell spiking response was not primarily dependent on total charge contained in the pulse train or current amplitude alone, but a combination of amplitude and frequency. Significance. These results indicate that HFS may be a promising method to target functionally-distinct neural pathways in the retina in an effort to improve the vision quality with retinal prostheses.

Electromyographic (EMG) is a bio-signal collected on human skeletal muscle. Analysis of EMG signals has been widely used to detect human movement intent, control various human-machine interfaces, diagnose neuromuscular diseases, and model neuromusculoskeletal system. With the advances of artificial intelligence and soft computing, many sophisticated techniques have been proposed for such purpose. Hybrid soft computing system (HSCS), the integration of these different techniques, aims to further improve the effectiveness, efficiency, and accuracy of EMG analysis. This paper reviews and compares key combinations of neural network, support vector machine, fuzzy logic, evolutionary computing, and swarm intelligence for EMG analysis. Our suggestions on the possible future development of HSCS in EMG analysis are also given in terms of basic soft computing techniques, further combination of these techniques, and their other applications in EMG analysis.