Search

Your search keyword '"Thornberry, Conor"' showing total 11 results
Start Over Author "Thornberry, Conor"
11 results on '"Thornberry, Conor"'

2. Understanding learning from EEG data: Combining machine learning and feature engineering based on hidden Markov models and mixed models

Author

Palma, Gabriel Rodrigues, Thornberry, Conor, Commins, Seán, and Moral, Rafael de Andrade

Subjects
Quantitative Biology - Quantitative Methods, Computer Science - Machine Learning, Electrical Engineering and Systems Science - Signal Processing
Abstract

Theta oscillations, ranging from 4-8 Hz, play a significant role in spatial learning and memory functions during navigation tasks. Frontal theta oscillations are thought to play an important role in spatial navigation and memory. Electroencephalography (EEG) datasets are very complex, making any changes in the neural signal related to behaviour difficult to interpret. However, multiple analytical methods are available to examine complex data structure, especially machine learning based techniques. These methods have shown high classification performance and the combination with feature engineering enhances the capability of these methods. This paper proposes using hidden Markov and linear mixed effects models to extract features from EEG data. Based on the engineered features obtained from frontal theta EEG data during a spatial navigation task in two key trials (first, last) and between two conditions (learner and non-learner), we analysed the performance of six machine learning methods (Polynomial Support Vector Machines, Non-linear Support Vector Machines, Random Forests, K-Nearest Neighbours, Ridge, and Deep Neural Networks) on classifying learner and non-learner participants. We also analysed how different standardisation methods used to pre-process the EEG data contribute to classification performance. We compared the classification performance of each trial with data gathered from the same subjects, including solely coordinate-based features, such as idle time and average speed. We found that more machine learning methods perform better classification using coordinate-based data. However, only deep neural networks achieved an area under the ROC curve higher than 80% using the theta EEG data alone. Our findings suggest that standardising the theta EEG data and using deep neural networks enhances the classification of learner and non-learner subjects in a spatial learning task., Comment: 25 pages

Published
2023

5. Frontal delta and theta power reflect strategy changes during human spatial memory retrieval in a virtual water maze task: an exploratory analysis.

Author

Thornberry, Conor and Commins, Sean

Subjects
*RECOLLECTION (Psychology), *SPATIAL memory, *TASK analysis, *MEMORY, *YOUNG adults, *COLLECTIVE memory, *ACTION theory (Psychology)
Abstract

Brain oscillations in humans play a role in a wide range of cognitive processes, including navigation and memory. The oscillatory dynamics contributing to successful spatial memory recall in humans are not well-understood. To investigate specific oscillatory frequency bands during the recall process in human navigation, we recorded electroencephalographic (EEG) activity during a recall trial in healthy young adults (n = 15) following the learning of a goal location in a Virtual Water Maze task. We compared this to the activity during the same trial length, in a group of participants who did not learn a target location and navigated freely but were time-matched to the learning group (non-learning, n = 15). We compared relative power in Delta (2-4Hz), Theta (5-7Hz), Alpha (8-12Hz), Beta (15-29Hz), and Gamma (30-40Hz) bands across the scalp. We found that delta and theta activity were greater during recall in our learning group, as opposed to our non-learning group. We also demonstrated clear suppression in the alpha band at posterior sites during memory-guided navigation compared to our non-learning group. Additionally, when goal-directed navigation switches to focused searching behavior, power becomes greater at the frontal region; with increases in the delta and theta bands reflecting this strategy change. There was also greater beta and gamma activity at posterior sites in our learning group. We discuss the results further in terms of the possible roles and functions of these oscillations during human navigation and hope this exploratory analysis can provide hypotheses for future spatial navigation and memory work. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

10. An Examination of the Influence of Landmarks in Human Spatial Navigation Using a Virtual Water Maze

Author

Thornberry, Conor

Abstract

Research into spatial memory and navigation excelled with the invention of the Morris Water Maze (Morris, 1984). In this task animals are required to find a platform, hidden somewhere in a large circular pool of water (below surface level). As animals cannot see the goal directly, they must use various cues in the environment to locate it and escape. Research has shown that landmarks exert control over an animal’s navigation ability. Recently, the Commins Lab has developed a virtual version of the Morris water maze task for use with humans; NavWell. This thesis established that the spatial behaviour of human participants navigating in NavWell is also controlled by virtual landmarks. In Experiment 1, participants trained to navigate with two landmarks, searched inaccurately during a recall trial with no landmarks and landmarks rotated 180°. However, does the visual saliency of these landmarks (e.g. brightness) influence our ability to recall a goal location during navigation? In Experiments 2, 3 & 4, we examined this question. In Experiment 2, participants were trained with a bright landmark near the target and a dim landmark far from the target. Participants were then examined with one cue in isolation or both. The group with the dim landmark searched incorrectly compared to the bright group. In Experiment 3, we controlled for brightness by switching the bright and dim landmark positions. Participants with the bright landmark searched incorrectly, but the dim group searched in the correct quadrant of the pool. In a final experiment, brightness was removed as a feature completely. Participants were trained with two landmarks of equal brightness levels. Here, the group with the landmark nearest the platform searched more accurately. The evidence for an associative learning model of human navigation, as well as the importance of proximity as a nontangible influence for landmark preference were then discussed.

Published
2019

Catalog

Books, media, physical & digital resources
See catalog results