Biomedical Engineering - Theses
Permanent URI for this collection
Search Results
1 - 2 of 2
-
ItemAdjusting the parameters of electrical stimulation of retinal ganglion cells to reduce neural adaptation and improve efficacy of retinal prostheses( 2018)Retinal prostheses aim to provide visual percepts through electrical stimulation of the retina to blind people affected by diseases caused by photoreceptor degeneration. Two challenges presented by current devices are a lack of selectivity in the activation of retinal ganglion cells (RGCs) and neural adaptation in the retina, which is believed to be the cause of fading—an effect where artificially produced percepts disappear over a short period of time, despite continuous stimulation of the retina. We aim to (1) understand the neural adaptation generated in RGCs during electrical stimulation, (2) obtain the preferred stimulation parameters (waveform) of each morphological class of RGCs and (3) use the preferred waveform of each morphological class to selectively activate different neurons. RGCs have been classified by morphology into 4 main groups: A, B, C and D. We performed an spike-triggered covariance (STC) analysis on the responses of 44 RGCs to extracellular electrical white noise and 43 RGCs to intracellular white noise. We then recovered their temporal electrical receptive fields (tERF), or waveform. A number of RGCs were stimulated with all the previously recovered waveforms to test the efficacy of each waveform on each other. The waveform recovered from the responses to intracellular stimulation have shown that RGCs can be classified into their respective morphological types by using a K-means clustering algorithm. Extracellular stimulation did not result in waveforms with a clear correlation between clusters and morphological classes. Cells from B and D morphological types had lower thresholds when stimulated with the waveform recovered from cells in the same morphological class. A-RGCs on the contrary, did not seem to share the same temporal features in their waveform with other A-type neurons. Further studies involving a larger data set might determine whether the waveform could preferentially stimulate cells from a specific morphological class. Current visual prostheses use electrical pulses with fixed frequencies and amplitudes modulated over hundreds of milliseconds to stimulate the retina. However, in nature, neuronal spiking occurs with stochastic timing, hence the information received naturally from other neurons by RGCs is irregularly timed. We used a single epiretinal electrode to stimulate and compare rat RGC responses to stimulus trains of biphasic pulses delivered at regular and random inter-pulse intervals (IPI), the latter taken from an exponential distribution. Our observations suggest that stimulation with random IPIs result in lower adaptation rates than stimulation with constant IPIs at frequencies of 50 Hz and 200 Hz. We also found a high proportion of lower amplitude action potentials, or spikelets. The spikelets were more prominent at high stimulation frequencies (50 Hz and 200 Hz) and were less susceptible to adaptation, but it was not clear if they propagated along the axon. Using random IPI stimulation in retinal prostheses reduces the decay of RGCs and this could potentially reduce fading of electrically induced visual perception.
-
ItemNeural tissue electrical modelling at micro and macro scales( 2018)A better understanding of electrical stimulation of the retina by neural prostheses may be essential for progress to be made towards a viable mass-market design of such devices. Dividing the problem into electrode models, target neuron models, and models of the tissue filling the volume between the electrodes and neurons, we focus on the tissue models. Prior work suggests that to model the relevant tissue, the neural retina, a standard, homogeneous, volume conductor may not be an appropriately faithful choice, even one with an anisotropic conductivity and permittivity. This is due to the capacitance of neural membranes and the macroscopic dimensions of the cable-like neural processes forming the tissue. Prior work on the subject resulted in alternative models being proposed (mean-field models). However, while those prior models may be solved approximately, there had been no well-established method to estimate the amount of error in those approximate solutions. We propose an alternative approach to derive a mean-field model, on the basis of finite element discretisations of a reference microstructural model. The latter is made up of infinitely-long axons running parallel to one another. To estimate the accuracy of those finite element solutions, we adapt the Global Convergence Index (GCI) technique. Our adaptation incorporates round-off error into the GCI technique in a systematic and conservative way. Our resulting mean-field model, the quantified-uncertainty (QU) model, produces solutions together with uncertainty estimates. While there are some differences between the QU model and prior models, they produce compatible solutions, in the sense that solutions using the prior models generally fall within the uncertainty band of solutions produced using the QU model, under boundary conditions of practical relevance. We describe a detailed method for solving a simple instance of a situated application problem incorporating the QU model. The derivation of the QU model proceeds by transforming the microstructural model into an appropriate spectral domain, then solving for a point source in a large, coarsely-discretised instance, in order to establish the claim that the far-field behaviour in two lattice directions is sufficient to characterise the whole response. We then solve finely-meshed finite element models corresponding to these two directions, under far-field boundary conditions. Observing that the solutions converge exponentially (and rapidly) towards functions with useful symmetry properties, we take advantage of the latter to constrain equivalent discrete models, reduced so as to represent only the quantities relevant to the mean-field description: potential, current flow across the fibres, and current flow along the fibres ("absorption"). We find equivalent continuous-domain models to the discrete models. We were also able to express the QU model in terms of two rational interpolating functions with a small number of coefficients. The uncertainty part of the QU model is formed so as to cover the differences between the two directions mentioned above, in addition to accounting for the fitting residuals from interpolation, for the discretisation error from the finite element representation and for the round-off error from solving the finite element matrices, including their conditioning.