In this research, we develop a neuromorphic system to study neural signaling at the level of first order tactile afferents which are slowly adapting type I (SA1) and rapidly adapting type I (RA1) mechanoreceptors. Considering, the linearized Izhikevich model, two digital circuits are developed for both afferents and are executed on the field programmable gate array (FPGA). After implementation of the digital circuits, we investigate how much information is encoded by this hardware-based neuromorphic system. Indeed, the artificial spiking sequences are evoked by applying different force profiles to the sensor connected to the FPGA. Next, the obtained neural responses are classified based on the two fundamental neural coding for brain information processing: spike timing and rate coding. Considering temporal coding, k-nearest neighbors (kNN), support vector machine (SVM) and Decision Tree algorithms are used for forces recognition using acquired artificial spike patterns. The results of classification show that the digital RA1 is susceptible to signal variations, while the digital SA1, on the other hand, is sensitive to the ramp and hold inputs. Furthermore, these responses are better distinguishable to different stimuli when both artificial SA1 and RA1 afferents are regarded. These results, which are functionally compatible with biological observations, yield the promise for fabrication and development of new tactile sensing modules to be employed in bio-robotic and prosthetic applications.

Spike train analysis in a digital neuromorphic system of cutaneous mechanoreceptor

Amiri M.;Falotico E.;Laschi C.
2020

Abstract

In this research, we develop a neuromorphic system to study neural signaling at the level of first order tactile afferents which are slowly adapting type I (SA1) and rapidly adapting type I (RA1) mechanoreceptors. Considering, the linearized Izhikevich model, two digital circuits are developed for both afferents and are executed on the field programmable gate array (FPGA). After implementation of the digital circuits, we investigate how much information is encoded by this hardware-based neuromorphic system. Indeed, the artificial spiking sequences are evoked by applying different force profiles to the sensor connected to the FPGA. Next, the obtained neural responses are classified based on the two fundamental neural coding for brain information processing: spike timing and rate coding. Considering temporal coding, k-nearest neighbors (kNN), support vector machine (SVM) and Decision Tree algorithms are used for forces recognition using acquired artificial spike patterns. The results of classification show that the digital RA1 is susceptible to signal variations, while the digital SA1, on the other hand, is sensitive to the ramp and hold inputs. Furthermore, these responses are better distinguishable to different stimuli when both artificial SA1 and RA1 afferents are regarded. These results, which are functionally compatible with biological observations, yield the promise for fabrication and development of new tactile sensing modules to be employed in bio-robotic and prosthetic applications.
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11382/531363
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
social impact