polito.it
Politecnico di Torino (logo)

Boosting Working Memory: Adaptive Neurostimulation in Virtual Reality

Gabriele D'Amato, Alice Fazio

Boosting Working Memory: Adaptive Neurostimulation in Virtual Reality.

Rel. Luca Mesin. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Biomedica, 2024

[img]
Preview
PDF (Tesi_di_laurea) - Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives.

Download (19MB) | Preview
Abstract:

Neurostimulation in closed-loop represents a promising technique for enhancing human cognitive functions. This thesis aims to explore its effectiveness in boosting the working memory (WM) of healthy individuals through virtual reality. The focus is on the improvement of specific aspects of WM such as the ability to maintain and manipulate information, sustained attention, and real-time information updating. These elements are crucial for a wide range of cognitive functions, directly impacting the efficient management of mental information, a critical aspect in daily life. Through a meticulous methodological approach, users are presented with an n-back task implemented in virtual reality with Unity, a platform that enables the development of interactive games. During the task, users' EEG signals are recorded using an 8-channel Enobio device and NIC2, software that controls the device from a computer. In the initial calibration phase, these tools operate offline. The recorded signal is then processed in Matlab, from which the most relevant features are extracted. These features, along with performance labels, are used to create a construction set for the training of an SVM regressor. In the subsequent real-time phase, users undergo 5 testing trials with conscious and unconscious stimuli, including neurostimulation with fixed or adaptive alpha-band Binaural Beats (BB), adaptive theta-band pulsed light, visual neurofeedback with a fill bar, and adaptive BB combined with neurofeedback. Additionally, a control condition without stimuli is included. During this phase, NIC2 transfers the signal to Matlab in real-time, where it is processed to calculate the Working Memory Level (WML) every 500 ms. Unity compares this index with personalized thresholds for the user and decides any changes to stimulation parameters to optimize performance. By recalculating subsequent WMLs, the positive or negative effect of this change is evaluated, thereby closing the loop. Task outcomes are assessed by considering the percentage of correct responses (PC), reaction time (RT), and the inverse efficiency score (IES), a metric that synthesizes the first two. In the 2-Back test, no significant differences are observed in the PC, RT, and IES parameters, possibly due to the test's simplicity, which limits errors, making it challenging to detect substantial improvements. Conversely, in the 3-Back test, significant differences emerge, with notable improvements in IES under alpha-band adaptive BB, both individually and in combination with neurofeedback (p < 0.05). The use of adaptive BB also lead to a significant reduction in RT compared to both the control conditions (p < 0.05) and the traditional use of constant BB (p < 0.05). Results suggest that alpha-band adaptive BB may contribute to the improvement of working memory performance in healthy subjects. Furthermore, they could prove beneficial for cognitive rehabilitation of pathological subjects, even in situations with limited conscious interactions, harnessing the power of 'unconscious brain entrainment'.

Relators: Luca Mesin
Academic year: 2023/24
Publication type: Electronic
Number of Pages: 148
Subjects:
Corso di laurea: Corso di laurea magistrale in Ingegneria Biomedica
Classe di laurea: New organization > Master science > LM-21 - BIOMEDICAL ENGINEERING
Aziende collaboratrici: Politecnico di Torino
URI: http://webthesis.biblio.polito.it/id/eprint/30494
Modify record (reserved for operators) Modify record (reserved for operators)