Selective Adaptation using Electrical Stimulation Devyani Nanduri 2, J. D. Weiland1,2, A. Horsager 1, M. S. Humayun 1, 2, R. J. Greenberg 4, M. J. McMahon 4, I. Fine 3 1Doheny Retina Institute, Doheny Eye Institute, Department of Ophthalmology, Keck School of Medicine, University of Southern California, Los Angeles, CA 2 Department of Biomedical Engineering, University of Southern California, Los Angeles, CA 3 Zilkha Neurogenetic Institute, Department of Psychology, University of Southern California, 4 Second Sight Medical Products Inc., Sylmar, CA Background Objective This experiment examined adaptation in the retina as a result of electrical stimulation using a retinal prosthesis. Goals 1. Measured how perceptual threshold of the retina changed with prior supra- threshold electrical stimulation 2. Compared adaptation effects for pulse trains of either short or long pulse width durations. Conclusions Weak evidence of adaptation selectivity for short pulse adaptors, which only affected short test pulses. This is consistent with short adaptors only adapting fast integrating cells, which mediate detection of short test pulses. In Subject 1, long pulse adaptors produced unselective effects consistent with long pulse widths adapting both fast and slow integrating cells. No adaptation for long pulse adaptors was found in Subject 2; higher current levels may be needed to observe adaptation effects. Biomedical Engineering Results Methods Each stimulus consisted of two consecutive trains of biphasic electrical pulses, each preceded by an auditory cue. The first adapting pulse train had a frequency of 45 Hz and a 1s duration. This adaptor was followed by a test stimulus of 15 Hz and 1s duration. A delay of 500 ms separated adaptor and test pulse trains. Adapting Stimuli at 45HzTest Stimuli at 15Hz 1 s 2.5s Auditory cue 1.5 s 0 Delay Time (s) Auditory cue Figure 2: Experimental Layout (1) No preceding adaptor (2) Short pulse width adaptor of 0.075ms (3) Long pulse width adaptor of 3ms. Support Data from 2 subjects, 6 electrodes and three experimental conditions were fit with strength duration curves using the power function y=a/x+b where the asymptote (b) was defined as the rheobase and the chronaxie was defined as a/b. Experiment Conditions Subject 1: Ratio of Adapt/No Adapt TEST Pulse Duration (ms) Adapt/No Adapt Subject 2: Ratio of Adapt/No Adapt TEST Pulse Duration (ms) Adapt/No Adapt Short Pulse Adapt Long Pulse Adapt No Adapt Short Pulse Adapt Long Pulse Adapt Pulse Duration (ms) Threshold (uA) Electrode Pulse Duration (ms) Threshold (uA) Electrode Pulse Duration (ms) Threshold (uA) Electrode Pulse Duration (ms) Threshold (uA) Electrode 2 Subject 1 Subject 2 Chronaxie values increased from no adapt (blue curves) to short pulse adapt (green dash-dot curves). Rheobase values increased from no adapt (blue curves) to long pulse adapt (red dash curves). Data did not show any significant changes in chronaxie and rheobase values between the different experimental conditions. Figure 5 shows adapt/no adapt for each subject. Figure 4 : Strength duration curves for two sample electrodes from two subjects Figure 5: Adapt/No Adapt for long and short pulse adapt conditions Retinitis pigmentosa (RP) and age-related macular degeneration are diseases where the photoreceptors progressively degenerate, while the cells of the inner retina remain relatively spared, though neurally disorganized (Marc, 2004). Since 2002, a collaborative effort between Second Sight Medical Products, Inc. and Doheny Eye Institute implanted six advanced RP patients with 4x4 epiretinal electrode arrays, with the goal of eliciting percepts by directly stimulating cells in the inner retina. The array of disk electrodes are implanted in the macular region and electrical signals are sent through an external inductive coupling device to an internal coil through a parallel system of wires to the epiretinal implant. Building a visual prosthetic requires us to understand the perceptual effects of continuous stimulation at both a neurophysiological and behavioral level. Electrical stimulation has the potential to stimulate a wide network of cells, including bipolar, amacrine and both on- and off-center ganglion cells. These various cell types integrate current with different time constants. For example, ganglion cells are thought to integrate current rapidly with a time constant of less than 1ms, whereas bipolar cells integrate current more slowly, with a time period of more than 10ms (Jensen, 2005). Here we examine whether adaptation might be selective for pulse width. If so, that might imply that pulse width can be used to differentially stimulate different cell types. Figure 1: Retinal Prosthesis with close-up of electrode array 1. Marc, R.E., Jones, B.W., Watt, C.B., & Strettoi, E. (2003). Neural remodeling in retinal degeneration. Progress in Retinal & Eye Research, 22 (5), Jensen, R.J., Ziv, O.R., Rizzo, J.F. (2005). Responses of rabbit retinal ganglion cells to electrical stimulation with an epiretinal electrode. Journal of Neural Engineering, 2(1):S Fried S.I., Hsueh H.A., Werblin F.S., (2006). A method for generating precise temporal patterns of retinal spiking using prosthetic stimulation. Journal of Neurophsyiology 95(2):970-8 Adapting stimuli were set at 2x threshold, as measured under no adapt conditions. Threshold was defined as the current amplitude needed to detect the stimulus on 50% of trials corrected for false alarms. Data were collected using a 1-up 1-down staircase with interleaved catch trials Threshold was measured for a 15 Hz, 1s test stimulus using pulse widths of ms, 0.45 ms and 3ms. We measured threshold for 3 conditions: Subject 1: Adapting trains of long pulses increased perceptual thresholds for short, medium and long test pulses. Adapting trains of short pulses only increased perceptual thresholds for short stimulus pulses. Subject 2: Slight adaptation effects for a short test pulse using a short pulse adaptor. In all other conditions, adaptation effects were insignificant. Long Pulse Adapt