Wavelength-sensitivity of mouse retinal ganglion cells recorded by a high-density micro electrode array (MEA)

Abstract

Event Abstract Back to Event Wavelength-sensitivity of mouse retinal ganglion cells recorded by a high-density micro electrode array (MEA) Florian Jetter1, Gabriel Bertotti2, Roland Thewes3 and Günther Zeck4* 1 NMI at the University Tuebingen, Neurochip Research, Germany 2 Technische Universität Berlin, Chair of Sensor and Actuator Systems, Germany 3 Technische Universität Berlin, Chair of Sensor and Actuator Systems, Germany 4 NMI at the University Tuebingen, Neurochip Research, Germany Motivation: High-density CMOS-MEAs can be used to simultaneously record the electrical spiking activity from hundreds of neurons [Bertotti et al., 2014]. Neuronal spiking can be induced by electrical stimulation [Eickenscheidt and Zeck 2014], by light stimulation applied to a light-sensitive retina [Zeck et al. 2011], or by light stimulation of optogenetically transfected neurons [Herrmann et al. 2014]. In this work, we investigate the wavelength-sensitivity of different mouse retinal ganglion cells and the intrinsic wavelength-sensitivity of the response of high-density CMOS-MEAs. Material and Methods: A CMOS-based high-density MEA comprising 4225 recording sites is used for recording the ganglion cell activity in C57/Bl6 mouse retina during flickering light stimulation (1 and 5 Hz, respectively) with different wavelength and of different light stimulus sizes. Light stimuli presented on a DMD (µ-matrix, Rapp Optoelectronic, Germany) are focused through a microscope objective on the retina. The DMD is illuminated by an LED system commonly used for optogenetic activation (pe-4000, coolLED, UK). Here the results for four stimulation wavelengths are presented (405, 470, 525, and 635nm) at intensities as high as 2 mW/mm² (470 nm). Recorded data is sampled at 25 kHz. Results: Light stimulation (stimulus area: 1 mm2) evokes spiking in the interfaced retina. Based on the stimulus polarity retinal ganglion cells are broadly classified in ON or OFF type, depending on whether they respond to light on- or offset, respectively. In Fig. 1A filtered recordings are shown of the measured extracellular voltage from two selected sensors during 1 Hz stimulation. Both, ON and OFF cell types respond to the three highest wavelengths used here. The sensor site recording the ON transient cell also detects activity from a second OFF cell with smaller amplitude. In Fig.1B we present the unfiltered extracellular voltage traces recorded by a third selected sensor site for all four wavelengths. Light onset leads to a measurable change of the current in the sensing transistor (cf. Bertotti et al., 2014) or more generally speaking of the sensed recording site response, reflected as a low-frequency change of signal back-converted into the voltage domain. However, this does not prevent the detection of light-induced spiking. We note, that the ganglion cell shown in Fig. 1B is not activated by the 635 nm light stimulus. Ongoing experiments investigate the sensitivity of retinal ganglion cells to chromatic stimulation and to stimuli presented at various intensities. The induced variations of the sensor signals are wavelength-dependent, with the highest change obtained for the lowest wavelength (405 nm) and undetectable changes for red light (635 nm). Conclusion: It is shown that recording of cellular spiking activity with CMOS-based MEAs is possible during optical stimulation. Ganglion cells in the mouse retina have different wavelength-sensitivities. The light-induced low-frequency drift of the sensor signals can be completely removed for the tested wavelengths and intensities using high-pass filtering.