Orientador: Prof. Dr. Marcio Eisencraft / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia da Informação, 2014. / Efficient techniques for analysis of neuronal signals in intracranial data acquisition systems
are almost prerequisites for the observation of specific brain regions for controlled stimulation,either for medical purposes or for human-computer interaction. The spike sorting, which aimsto classify firings per neuron, is relevant to simplify the analysis of noisy neuronal signalsthat possess a significant amount of information. Detection is the first step of the spikesorting and is decisive in terms of efficiency. In this case, what makes the detection problemrather difficult is the fact that we do not know exactly the waveform of the signal to bedetected and the firing instants. In the present work, the limits of detection techniques
based on threshold are studied considering a simplified model of neuronal signals. The
techniques are evaluated using error probability measures as a function of the signal-to-noiseratio (SNR), of the firing rate and of the threshold level used. The objective is to combine theadvantages and limits of good performance of detection techniques with the characteristicsof different time series considered. Computational simulations suggest that: (i)the SNR andthe interval between firings significantly affect the relative performance of the techniques;(ii) the optimal threshold is a function of the statistical characteristics of the time seriesand, if properly chosen, can avoid detection errors in some situations and (iii) the thresholdcalculations proposed in the literature may not be the most effective. The results of thesimplified model should serve as an estimate of the worst case performance for each methodwhen they are applied in situations with acquired real data.
| Identifer | oai:union.ndltd.org:IBICT/oai:BDTD:76617 |
| Date | January 2014 |
| Creators | Castellani, Renato Aguiar |
| Contributors | Eisencraft, Marcio, Suyama, Ricardo, Paula, Amanda Souza de |
| Source Sets | IBICT Brazilian ETDs |
| Language | Portuguese |
| Detected Language | English |
| Type | info:eu-repo/semantics/publishedVersion, info:eu-repo/semantics/masterThesis |
| Format | application/pdf, 88 f. : il. |
| Source | reponame:Repositório Institucional da UFABC, instname:Universidade Federal do ABC, instacron:UFABC |
| Rights | info:eu-repo/semantics/openAccess |
| Relation | http://biblioteca.ufabc.edu.br/index.php?codigo_sophia=76617&midiaext=69790, http://biblioteca.ufabc.edu.br/index.php?codigo_sophia=76617&midiaext=69789, Cover: http://biblioteca.ufabc.edu.brphp/capa.php?obra=76617 |
Page generated in 0.0017 seconds