Title :
Recreating oscillatory behavior in artificial nerve cells
Author :
Wolpert, Seth ; Chinwalla, Alifya E.
Author_Institution :
Dept. of Electr. Eng., Maine Univ., Orono, ME, USA
Abstract :
The recreation of the phenomenon of cyclic inhibition between two artificial nerve cells is discussed. Implemented by means of postinhibitory rebound, it was achieved in a mutually inhibiting pair of identical nerve cell circuits, or neuromimes. Each neuromime is nominally self-excitatory, with a single inhibitory input, derived from the output of its counterpart. Postinhibitory rebound is accomplished by dynamically varying the synaptic weight of that inhibitory input. The neuromime employed is a comprehensive VLSI-based circuit with a host of features attributable to biological nerve cells. It was fabricated in two-micron CMOS technology with a total chip area of 0.6 square millimeters, and requires only a few passive discrete components for support and adjustment
Keywords :
CMOS analogue integrated circuits; VLSI; learning (artificial intelligence); neural chips; neurocontrollers; CMOS; VLSI-based circuit; Widlar source; artificial nerve cells; cyclic inhibition; mutually inhibiting pair; neuromimes; oscillatory behavior; passive discrete components; postinhibitory rebound; self-excitatory; synaptic weight; two-micron; Biomembranes; CMOS technology; Cells (biology); Circuit testing; Electron tubes; Neurons; Oscillators; Shape control; Vacuum technology; Very large scale integration;
Conference_Titel :
Bioengineering Conference, 1993., Proceedings of the 1993 IEEE Nineteenth Annual Northeast
Conference_Location :
Newark, NJ
Print_ISBN :
0-7803-0925-1
DOI :
10.1109/NEBC.1993.404398
