ETD

Archivio digitale delle tesi discusse presso l'Università di Pisa

Tesi etd-01252022-022955


Tipo di tesi
Tesi di laurea magistrale
Autore
de ALTERIIS, GIUSEPPE
URN
etd-01252022-022955
Titolo
Analysis of Neuropixels Data to understand the effects of Sleep Deprivation
Dipartimento
INGEGNERIA DELL'INFORMAZIONE
Corso di studi
BIONICS ENGINEERING
Relatori
relatore Prof. Oddo, Calogero Maria
relatore Prof.ssa Cirelli, Chiara
relatore Prof. Tononi, Giulio
Parole chiave
  • Computational Neuroscience
  • Sleep
  • Spike Sorting
Data inizio appello
11/02/2022
Consultabilità
Non consultabile
Data di rilascio
11/02/2092
Riassunto
This dissertation is the result of my experience working at the Center for Sleep and Consciousness at the University of Wisconsin-Madison. I collaborated with a group of graduate students and post-doctoral fellows whose general aim is to use and develop tools for electrophysiological data analysis to understand why we need to sleep. One way to address this question is to record the activity of neurons, and see how they behave during waking, sleep, and after sleep deprivation. In my project we used high quality chronic Neuropixels recordings in rats. Neuropixels probes are a recent innovation in Neural Engineering and allow to record the activity of hundreds or thousands of single neurons (single units). Our specific research question was to assess whether OFF periods during sleep are homeostatically regulated. OFF periods are short epochs (less than a second) during which neurons stop firing; they always occur during sleep, but our question was whether they get longer after sleep deprivation. If so, this would indicate that neurons get “tired” by firing during waking, and need to stop firing for longer periods during sleep when waking is prolonged beyond its physiological duration. Technically speaking, detecting and measuring OFF periods at the single unit level is challenging. One main reason is the significant drift of the probes that can occur in the brain tissue, especially in chronic recordings. Because of the drift, it is difficult to be sure that a specific channel is always detecting the spikes belonging to the same unit. I contributed to the computational part of the project, and to develop a processing pipeline for chronic Neuropixels analyses in the lab. The pipeline consists of preprocessing, spike sorting, postprocessing and curation of the output.

Introduction: The First Chapter of the thesis presents its relevance both from a Neural Engineering and Neuroscience perspective. It discusses some of the most important questions in sleep research and describes the sleep slow oscillation, whose correlates at the cellular level are ON and OFF periods. Moreover, it presents the technological advancements of Neuropixels probes and finally the rationale of the work. Materials and Methods: The Second Chapter describes the experiments and all the processing algorithms. The Third Chapter explains our use of the Phy software for the manual curation of the sorted output. It is structured as a tutorial. Finally, it describes the Data Analysis steps to select well isolated single units and discard noise-contaminated ones, to evaluate them based on quality metrics, and perform the statistical comparisons. Results: The Fourth Chapter presents the technical results and the biological results. The technical results section Section shows the improvements that have been introduced in the spike sorting pipeline. The biological results Section show that OFF periods get longer after sleep deprivation.
Discussion and Conclusion: Finally, the Fitfh Chapter discusses both the technical results and the biological results. The improvements of the sorting pipeline can be applied in several chronic Neuropixels experiments, even not regarding sleep. Moreover, the drift correction algorithm can be potentially employed in a wide range of neural probes. However, there are still some open questions, such as how to reduce the number of noise contaminated clusters identified by the spike sorting algorithm. The biological results Section suggests that neurons need a bigger rest period after sleep deprivation. The mechanism by which this happens is still uncertain.

File