MRS Meetings and Events

 

EL04.04.04 2023 MRS Fall Meeting

An Organic Spiking Neuron with Unconventional Form-Factor for Energy Efficient, Neurotransmitter-Mediated, In-Sensor Pre-Processing Functions

When and Where

Nov 29, 2023
9:30am - 9:45am

Hynes, Level 3, Room 313

Presenter

Co-Author(s)

Giovanni Maria Matrone1,Xudong Ji1,Abhijith Surendran1,Zachary Laswick1,Gang Ye2,Francesca Santoro3,Yoeri van de Burgt4,Jonathan Rivnay1

Northwestern University1,Shenzhen University2,Forschungszentrum Juelich3,Technische Universiteit Eindhoven4

Abstract

Giovanni Maria Matrone1,Xudong Ji1,Abhijith Surendran1,Zachary Laswick1,Gang Ye2,Francesca Santoro3,Yoeri van de Burgt4,Jonathan Rivnay1

Northwestern University1,Shenzhen University2,Forschungszentrum Juelich3,Technische Universiteit Eindhoven4
The fundamental mechanisms of signal communication within the human body rely on the spiking frequency of action potentials.<sup>1</sup> Through biological receptors and afferent neuronal cells, stimuli from the external world are encoded into a spiking pattern and transmitted to the central nervous systems where they are processed via interneurons: the “sensory coding” mechanisms.<br/>The fundamental goal of neuromorphic electronics is to emulate the architecture of the human brain to enable parallel computing with high energy efficiency<sup>2</sup> and local processing, thus advancing intelligent systems interfacing with the human body<sup>3</sup>.<br/>Recently, organic materials have been employed to build electronic circuits that mimic both the spiking behaviour of neurons<sup>4–6</sup> and their synaptic transmission<sup>7</sup> thus replicating afferent neurons and interneurons functions. Indeed, a combination of bioelectronic devices may recreate a “neuronal pathway” that in nature relies on the cooperation of spiking (neurons and interneurons) and non-spiking elements such as mechano-chemical sensors (receptors), and neuromodulator junctions (chemical synapses)<sup>8</sup>. Although the most recent neuromorphic circuits emulate biological functions which are deemed essential for basic signal processing and computation capabilities (mimicking some retinal functions), i) these systems power consumption is 2-3 order of magnitude higher than the brain processors, ii) their footprint is still on a scale (tens of mm) impeding meaningful in-sensor applications, iii) the electrochemical detectors for neurotransmitter still show sensitivity and selectivity not suitable for tissue interfacing.<br/>Here it is presented an integrated neuromorphic platform that replicates both afferent neurons “sensory coding” and synaptic transmission.<br/>Exploiting a vertical transistor architecture comprising stacked n-p type polymer films, complementary inverters are fabricated with a lateral footprint of 50 mm and requiring low drain voltage (0.05 V), thus building an independent spiking unit. To allow neurotransmitter-mediated synaptic transmission with high sensitivity, a novel biohybrid synapse is developed exploiting a referenced-ENODe architecture<sup>9</sup>. This system allows the detection of neurotransmitters, such as dopamine and serotonin, with a sensitivity approaching sub-nanomolar concentrations and enhancement selectivity, facilitating tissue coupling and potentially avoiding molecular cross-talk.<br/>Moreover, the total neuromorphic system energy consumption, as predicted by electronic circuit simulations, is reduced to 1 nJ<sup>10</sup>, with a synaptic transmission contribution depending on the device lateral dimension and an energy-per-spike depending on the materials selection and the dimension of the vertical inverters. Moreover, the integrate-and-fire neuron footprint is reduced to 200mm x 200mm. This system constitutes a fundamental building block for programmable neural pathways. It is compatible with in-sensor application footprint and sensitivity requirements for locally executing bio-inspired pre-processing functions while its unconventional architecture allows to dynamically communicate with the nervous system.<br/><br/><b>1</b> Kandel, E. R. <i>et al.</i> 4, (McGraw-hill New York, 2000)<br/><b>2</b> Furber, S. <i>J. Neural Eng.</i> 13, 051001 (2016)<br/><b>3</b> Yoo, J. <i>et al.</i> <i>Current Opinion in Biotechnology</i> 72, 95–101 (2021)<br/><b>4</b> Mirshojaeian Hosseini, M. J. <i>et al.</i> <i>J. Phys. D: Appl. Phys.</i> 54, 104004 (2021)<br/><b>5</b> Harikesh, P. C. <i>et al.</i> <i>Nat Commun</i> 13, 901 (2022)<br/><b>6</b> Harikesh, P. C. <i>et al.</i> <i>Nat. Mater.</i> 22, 242–248 (2023)<br/><b>7</b> Matrone, G. M. <i>et al.</i> <i>Adv Materials Technologies</i> 2201911 (2023)<br/><b>8</b> Matrone, G. M. <i>et al.</i> (In Review, 2022)<br/><b>9</b> Ji, X. <i>et al.</i> <i>Nat Commun</i> 14, 1665 (2023)<br/><b>10</b> Lee, Y. <i>et al.</i> <i>Joule</i> 5, 794–810 (2021)

Symposium Organizers

Simone Fabiano, Linkoping University
Paschalis Gkoupidenis, Max Planck Institute
Zeinab Jahed, University of California, San Diego
Francesca Santoro, Forschungszentrum Jülich/RWTH Aachen University

Symposium Support

Bronze
Kepler Computing

Publishing Alliance

MRS publishes with Springer Nature