Title Oblikovanje neuromodulacijskog pulsa optičkom pobudom organskih elektrolitskih foto-kondenzatora
Title (english) Shaping the neuromodulation pulse by optical stimulation of organic electrolytic photocapacitors
Author Marta Nikić
Mentor Vedran Đerek (mentor)
Committee member Vedran Đerek (predsjednik povjerenstva)
Committee member Davor Horvatić (član povjerenstva)
Committee member Željko Skoko (član povjerenstva)
Committee member Mihael Makek (član povjerenstva)
Committee member Nenad Pavin (član povjerenstva)
Granter University of Zagreb Faculty of Science (Department of Physics) Zagreb
Defense date and country 2021-09-03, Croatia
Scientific / art field, discipline and subdiscipline NATURAL SCIENCES Physics
Abstract Poznavanje ponašanja uređaja i elektroda predviđenih za implantaciju i stimulaciju u biološkim organizmima radi znanstveno istraživačkih i terapijskih primjena preduvjet je za njihovo korištenje. U sklopu ovog diplomskog rada istraživao se odziv optoelektroničkog uređaja predviđenog za stimulaciju živčanih stanica pod imenom foto-kondenzator. Dizajniran u sklopu grupe “Organski nanokristali” koja djeluje u okviru Laboratorija za organsku elektroniku (LOE) na Sveučilištu u Linkopingu (LiU), sastoji se od 3 tanka sloja, metala i p-n dvosloja dva organska poluvodiča. Prilikom njegovog obasjavanja svjetlošću dolazi do njegovog nabijanja koje uzrokuje lokalne struje u elektrolitu. Te lokalne struje u elektrolitu su posljedica gradijenta potencijala oko foto-kondenzatora koji ima ključnu ulogu prilikom stimulacije stanica. Promjenom oblika svjetlosnog pulsa želio se ispitati odziv foto-kondenzatora. Kako bi se navedeno moglo i predvidjeti, pristupilo se analitičkim, numeričkim i metodama rješavanja problema upotrebom strojnog učenja. Također, razmatran je inverzni problem u kojem se iz poznavanja odziva foto-kondenzatora želi predvidjeti potreban oblik svjetlosnog pulsa kojim se on može postići. Analitički pristup zbog nelinernosti sustava diferencijalnih jednadžbi nije detaljno obrađivan, međutim numeričkim pristupom koji koristi program za simulaciju strujnih krugova LTSpice određeni su parametri ekvivalentnog strujnog kruga sustava foto-kondenzatora i elektrolita, te su omogućena predviđanja odziva foto-kondenzatora za dani oblik svjetlosnog pulsa. Kako bi se rezultati predviđanja poboljšali i kako bi se mogao riješiti i inverzni problem, pristupilo se korištenju metode strojnog učenja. Izrađena je arhitektura modela koja je istrenirana i testirana s ukupno 15049 parova signala oblika strujnog pulsa i odziva foto-kondenzatora. Snimljeni signali su prikupljeni uz pomoć koda u programskom jeziku Python i automatskog prikupljanja podataka sa USB osciloskopa (Picoscope model 4424A). Python kod uključuje automatizaciju prikupljanja izmjerenih signala i promjene oblika svjetlosnog pulsa koji je s tehničke strane postignut generatorom signala Picoscopea. Uz pomoć dubokog učenja, izrađen je model koji nakon treniranja može predviđati odzive foto-kondenzatora za dani oblik svjetlosnog pulsa, te isti model također može riješiti inverzni problem, odnosno predvidjeti potrebni oblik svjetlosnog pulsa uz dani odziv foto-kondenzatora.
Abstract (english) Knowing the way how devices and electrodes meant for stimulation in biological systems work is essential for their usage. In this master thesis work the response of an optoelectronic device intended for neurostimulation is explored. This device, named a photocapacitor, designed by Organic Nanocrystal group at the University of Linkoping (LiU), is made of 3 thin film layers, metal and p-n junction of the two organic semiconductors. When shined by light, it charges and causes local currents in the electrolyte that are product of gradient of the potential around photocapacitor which has the key role during the stimulation of the cell. With changing the shape of the light pulse used, the effect on response of photocapacitor was intended to be observed. For the said to be predicted, analytical, numerical and deep learning methods were employed. Also, the inverse problem, in which the response of the photocapacitor is known and the needed shape of the light pulse is wanted, was considered. Analytical method due to the non-linearity of the system of differential equations was not tackled in details, but with the numerical method used by software for simulation of electric circuits, LTSpice, parameters of equivalent electric circuit of the system of photocapacitor and electrolyte were determined. Consequently, predictions of the response of the photocapacitor were possible. To make the results of the predictions more accurate and to be able to solve the inverse problem, a deep learning method was employed. Architecture of a model was trained and tested using 15049 pairs of signals of the shapes of light pulses and responses of photocapacitor. Measured signals were collected using Python code and a USB oscilloscope (Picoscope model 4424A). Python code delivers automation of collecting of measured signals and changing of shapes of light pulses. Light pulses were modulated using a signal generator embedded within the Picoscope. Model was trained and was successful in predicting the response of the photocapacitor for given shape of the light pulse. Furthermore, the same model was also successfully used to solve inverse problem.
Keywords
foto-kondenzator
bioelektronika
LTSpice
duboko učenje
Keywords (english)
photocapacitor
bioelectronics
LTSpice
deep learning
Language croatian
URN:NBN urn:nbn:hr:217:082113
Study programme Title: Physics; specializations in: Research Course: Research Study programme type: university Study level: integrated undergraduate and graduate Academic / professional title: magistar/magistra fizike (magistar/magistra fizike)
Type of resource Text
File origin Born digital
Access conditions Open access
Terms of use
Created on 2021-11-12 10:08:43