Revealing Structures in Data: Understanding Innate Behaviors and Neural Representations Using Latent Variables

Horno, Oihane

http://hdl.handle.net/10362/170658

Utilize este identificador para referenciar este registo.

Nome:	Descrição:	Tamanho:	Formato:
OihaneHorno_Thesis_Final_Corrected.pdf		16.69 MB	Adobe PDF	Ver/Abrir

Contacte-nos

Autores

Horno, Oihane

Orientador(es)

Machens, Christian

Petreanu, Leopoldo

Lima, Susana

Resumo(s)

"This thesis comprises two distinct projects. The first project employs Principal Component Analysis (PCA), a dimensionality reduction technique, to investigate how a population of neurons in the mouse primary visual cortex (V1) represents visual stimuli in the presence and absence of direct cortico-cortical Feedback (FB) from the Higher Visual Areas (HVA). Specifically, we examine the representation of two distinct stimuli: drifting gratings and a naturalistic movie. Our findings demonstrate that changes in representation are stimulus specific. In the case of drifting gratings, the absence of feedback leads to an increase in population gain, while for the naturalistic movie, it alters the representational geometry. These results support the notion that feedback provides contextual information to V1 neurons. (...)"