Dynamics of non-standard retinal-sensor behavior for natural images processing.
Objective: Collect existing and new experimental facts on non-standard behaviors of retinal ganglions cells in order to understand the computational capabilities of the retina under natural dynamic simulations.
Methods: Multielectrodes arrays (MEA) system (64 X PLEXON amplifier, National Instrument (NI625X) acquisition board (10-25 KHz) and hard drives for the storage) for field potential or spike recording is operational in our lab. MEA arrays are commercially available (Multichannel System).
The retina is mounted in a perfusion camera with AMES or equivalent solution, oxygenates and maintained at 35oC, under controlled light conditions. Labview homemade software (SpikeHunter) for signal acquisition, control and exportation and visualization (SpikePlay) is use. OFF line visualization, analysis, spike clustering and cross-correlograms response are done with MATLAB routines and following (Schwartz and Berry, 2008; Schwartz et al., 2007) protocols. An advantage of recording from as many 50-100 using MEA arrays, neurons is the simultaneously access to “population coding” from a complete retina ganglion cell assembly. The physiological characteristics of ganglion cells with respect to natural images stimulus, including (e.g.) contrast-intensity, color, texture, will be uses to characterize properties of firings (transients, sustain) ON / OFF, size and organization of receptive fields. The Universidad de Valparaiso is an institution with an approved animal welfare assurance (A5823-01) from the NIH (USA).
Figure joint shows our protocol and data samples. Inset A) Labview home made software for MEA data recording. We see some spike activity in some of the channels; B) MEA matrix 64 X electrodes; C) SpikeIt homemade software e.g. two acquisition channels and in the bottom the sequence of light flashes. One channel shows spikes; D) a plot raw data with more channels and spikes (dark vertical activity); E) As example, two templates for spike sorting base on data in D) and F) e.g. raster of response for 3 different ganglion cells (the plot alignment is displace compare to D).
Task steps:
(i) Review of available experimental data on retinal responses to dynamic and natural stimuli, the proposed models and the emerging computational properties (this preliminary task is important in this multidisciplinary context).
(ii) Design of the new stimulation protocols and validation of the natural image sequence stimulator.
(iii) Physiological experiments recording the retinal activity, using Patch clamp and Multielectrodes, methods, in rodents and amphibians, in response to natural (e.g. movies) visual stimulus.
(iv) Data basing and preprocessing of the ganglion cells recordings.
