Electrical and Electronic Engineering - Research Publications

Permanent URI for this collection

http://hdl.handle.net/11343/362

Filters

2014 - 2015 2
2
Reset filters

Settings
Statistics
Citations
Author: Burkitt, Anthony × Author: Grayden, David × Author: Meffin, Hamish × Author: Nesic, Dragan ×

Search Results

Now showing 1 - 2 of 2
  • Item
    Thumbnail Image
    A comparison of open-loop and closed-loop stimulation strategies to control excitation of retinal ganglion cells
    Kameneva, T ; Zarelli, D ; Nesic, D ; Grayden, DB ; Burkitt, AN ; Meffin, H (Elsevier, 2014-11-01)
    Currently, open-loop stimulation strategies are prevalent in medical bionic devices. These strategies involve setting electrical stimulation that does not change in response to neural activity. We investigate through simulation the advantages of using a closed-loop strategy that sets stimulation level based on continuous measurement of the level of neural activity. We propose a model-based controller design to control activation of retinal neurons. To deal with the lack of controllability and observability of the whole system, we use Kalman decomposition and control only the controllable and observable part. We show that the closed-loop controller performs better than the open-loop controller when perturbations are introduced into the system. We envisage that our work will give rise to more investigations of the closed-loop techniques in basic neuroscience research and in clinical applications of medical bionics.
  • Item
    Thumbnail Image
    Spike history neural response model
    Kameneva, T ; Abramian, M ; Zarelli, D ; Nesic, D ; Burkitt, AN ; Meffin, H ; Grayden, DB (SPRINGER, 2015-06)
    There is a potential for improved efficacy of neural stimulation if stimulation levels can be modified dynamically based on the responses of neural tissue in real time. A neural model is developed that describes the response of neurons to electrical stimulation and that is suitable for feedback control neuroprosthetic stimulation. Experimental data from NZ white rabbit retinae is used with a data-driven technique to model neural dynamics. The linear-nonlinear approach is adapted to incorporate spike history and to predict the neural response of ganglion cells to electrical stimulation. To validate the fitness of the model, the penalty term is calculated based on the time difference between each simulated spike and the closest spike in time in the experimentally recorded train. The proposed model is able to robustly predict experimentally observed spike trains.