Tesi etd-03302023-172334
Link copiato negli appunti
Tipo di tesi
Dottorato
Autore
FRUZZETTI, LORENZO
URN
etd-03302023-172334
Titolo
Neural-Inspired Motor Control and Adaptation: From Biological Systems to Soft Robotics
Settore scientifico disciplinare
ING-IND/34
Corso di studi
Istituto di Biorobotica - PHD IN BIOROBOTICA
Commissione
relatore FALOTICO, EGIDIO
Relatore Prof. CIANCHETTI, MATTEO
Membro Prof. GARRIDO, JESUS
Relatore Prof. CIANCHETTI, MATTEO
Membro Prof. GARRIDO, JESUS
Parole chiave
- Nessuna parola chiave trovata
Data inizio appello
30/11/2023;
Disponibilità
completa
Riassunto analitico
This research delves into the intricate microcircuitry and modeling of the motor system. The motivation for this exploration stems from several grounds: the biological accuracy of neural circuitry, its role in capturing temporal dynamics, applications in robotics and neuroprosthetics, potential benefits for rehabilitation and therapy, predictive modeling capabilities, insights into brain diseases, and understanding neural plasticity and learning.
While it's understood that different brain regions perform distinct computations, the specifics of these computations remain largely ambiguous. Particularly, the mechanisms through which circuits of interconnected neurons process information and the refinement of movements based on bodily feedback are yet to be deciphered. By studying motor system circuitry, we aim to advance control algorithms.
Our approach incorporated two main studies:
- A neurophysiological analysis in the mouse premotor cortex focused on recording individual units. We classified these units into putative Fast Spiking Neurons (FSNs) and Pyramidal Neurons (PNs) based on waveform characteristics, aiming to discern how information flows between these classes and what they predominantly encode.
- comprehensive spiking model of the cerebellum was crafted to understand temporal information flow, especially regarding how plasticity influences cerebellar output.
Our insights into information flow within movement-associated areas provide clarity on questions related to the temporal encoding of the movement in the premotor cortex.
Findings from the mouse premotor cortex indicated that FSNs displayed a more prolonged firing duration than PNs during licking, but not for forelimb movements. Our computational analysis established that FSNs held significantly more data about movement onset, however the information content of them is more reduntant. Meanwhile, the cerebellar model identified the roles of LTP and LTD in refining saccade kinematics, with LTD enhancing accuracy by reducing visual error and LTP amplifying peak speed.
Looking ahead, future avenues include a deeper neurophysiological exploration by identifying specific classes, their interactions, and the potential for optogenetic modulation. Additionally, studying plasticity across the motor system and discerning common organization paradigms will offer more comprehensive insights into movement generation and control.
While it's understood that different brain regions perform distinct computations, the specifics of these computations remain largely ambiguous. Particularly, the mechanisms through which circuits of interconnected neurons process information and the refinement of movements based on bodily feedback are yet to be deciphered. By studying motor system circuitry, we aim to advance control algorithms.
Our approach incorporated two main studies:
- A neurophysiological analysis in the mouse premotor cortex focused on recording individual units. We classified these units into putative Fast Spiking Neurons (FSNs) and Pyramidal Neurons (PNs) based on waveform characteristics, aiming to discern how information flows between these classes and what they predominantly encode.
- comprehensive spiking model of the cerebellum was crafted to understand temporal information flow, especially regarding how plasticity influences cerebellar output.
Our insights into information flow within movement-associated areas provide clarity on questions related to the temporal encoding of the movement in the premotor cortex.
Findings from the mouse premotor cortex indicated that FSNs displayed a more prolonged firing duration than PNs during licking, but not for forelimb movements. Our computational analysis established that FSNs held significantly more data about movement onset, however the information content of them is more reduntant. Meanwhile, the cerebellar model identified the roles of LTP and LTD in refining saccade kinematics, with LTD enhancing accuracy by reducing visual error and LTP amplifying peak speed.
Looking ahead, future avenues include a deeper neurophysiological exploration by identifying specific classes, their interactions, and the potential for optogenetic modulation. Additionally, studying plasticity across the motor system and discerning common organization paradigms will offer more comprehensive insights into movement generation and control.
File
| Nome file | Dimensione |
|---|---|
| Thesis_L...zetti.pdf | 5.02 Mb |
Contatta l'autore |
|