Harmony in early visual cortex: uncovering the complementary roles of gamma & alpha oscillations in local & global processing

Duecker, Katharina ORCID: 0000-0001-7887-3395 (2024). Harmony in early visual cortex: uncovering the complementary roles of gamma & alpha oscillations in local & global processing. University of Birmingham. Ph.D.

Preview

Duecker2024PhD.pdf
Text - Accepted Version
Available under License Creative Commons Attribution Non-commercial.
Download (8MB) | Preview

Abstract

Upon opening our eyes, we immediately perceive the world around us. Our remarkably seamless visual perception is a major source of information when interacting with our surroundings. Indeed, the dominance of vision over other senses is so significant that the brain allocates more resources to it than to any other sensory modality (Colavita, 1974; Van Essen et al., 1992). Yet, our seemingly effortless visual experience relies on a complex network of intricately interconnected cortical and sub-cortical structures performing real-time computations.

A central question in neuroscience is how the visual system optimally processes and routes incoming information. Popular ideas posit that neuronal oscillations, specifically gamma (> 30 Hz) and alpha oscillations (8 - 12 Hz), coordinate visual processing and attention (Fries, 2005, 2015; Jensen, 2023; Jensen & Mazaheri, 2010; Klimesch, 2012; Klimesch et al., 2007).

In this thesis, I will present findings based on empirical and computational work showing how gamma and alpha oscillations modulate visual inputs. The empirical part of my thesis consists of two studies using Magnetoencephalography (MEG) in combination with a high-frequency (subliminal) visual flicker. I will reveal that gamma oscillations in early visual regions are robust against this external, high-frequency stimulation and do not synchronise to the visual flicker. As I will discuss, this finding challenges the prevalent notion that gamma oscillations are critical for inter-areal communication in the visual system (also see Schneider et al., 2021, 2023; Vinck et al., 2023).

Following my experience with high-frequency stimulation gained in the first empirical chapter, I will use Rapid Invisible Frequency Tagging (RIFT) in the third chapter to investigate the neural correlates of feature-guided visual search. RIFT is a novel, subliminal stimulation technique used to probe cortical excitability to visual inputs (Zhigalov et al., 2019). I will demonstrate that alpha oscillations globally modulate the cortical excitability to the visual search display, which was further linked to improved search performance. This suggests that gain modulation by alpha oscillations can support performance in visual attention tasks with a high number of distracting stimuli. However, I will also discuss a set of analyses suggesting that the response time modulation by the alpha rhythm may be linked to task duration. I will offer different perspectives on these multi-faceted results, as well as future research directions to understand the relationship between inhibition by alpha oscillations and visual search performance.

Inspired by these empirical results, I will finally present a dynamical artificial neural network — a computer vision algorithm embracing the rhythmic dynamics of the visual cortex. I will demonstrate that this network, despite not being explicitly trained for the task, can handle multiple concurrent visual inputs by segregating them in time.

In summary, my thesis combines empirical and computational methods to explore how gamma and alpha oscillations contribute to computational processes in the visual system. I will conclude that while gamma oscillations reflect localised neural processes, alpha oscillations operate at a more global scale.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):