Optimal decoding of spatiotemporal patterns in Magnetoencephalography (MEG)
Lead Research Organisation:
Cardiff University
Department Name: Sch of Psychology
Abstract
Acting based on intention is a fundamental ability to our lives. Apple or orange, cash or card: we constantly make intentional decisions to fulfill our desires, even when the options have no explicit difference in their rewards.
This PhD project will combine brain imaging at different modalities (functional and structural MRI, MR spectroscopy and Magnetoencephalography) with computational modelling to understand how the human brain makes intentional decision, and how intentional decisions differ between and within individuals. This project will focus on healthy individuals. In the long run, results from this project may have implications to the mechanistic understanding of decision deficits in patients with neurodegenerative diseases. The studentship links to a European Research Council grant held by Dr Jiaxiang Zhang at the Cardiff University Brain Research Imaging Centre (CUBRIC). The student on this project would receive multidisciplinary training in cognitive neuroscience, brain imaging and computational modelling. We encourage students from different disciplines (e.g., cognitive neuroscience, imaging
neuroscience, psychology, computer science, mathematics or physics)
This PhD project will combine brain imaging at different modalities (functional and structural MRI, MR spectroscopy and Magnetoencephalography) with computational modelling to understand how the human brain makes intentional decision, and how intentional decisions differ between and within individuals. This project will focus on healthy individuals. In the long run, results from this project may have implications to the mechanistic understanding of decision deficits in patients with neurodegenerative diseases. The studentship links to a European Research Council grant held by Dr Jiaxiang Zhang at the Cardiff University Brain Research Imaging Centre (CUBRIC). The student on this project would receive multidisciplinary training in cognitive neuroscience, brain imaging and computational modelling. We encourage students from different disciplines (e.g., cognitive neuroscience, imaging
neuroscience, psychology, computer science, mathematics or physics)
Organisations
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509449/1 | 01/10/2016 | 30/09/2021 | |||
| 1982622 | Studentship | EP/N509449/1 | 01/10/2017 | 31/03/2021 | Dominik Krzeminski |
| Description | The aim of this project was to develop new methods for the brain's MEG signal decoding. We applied a modified method from statistical physics (Ising model) to resting-state MEG recordings. Thanks to that we could analyse statistical properties of signal co-activation between various oscillatory patterns and resting-state networks. To test the usefulness of our approach, we used a dataset comprising 26 recordings of healthy subjects and 26 patients suffering from juvenile myoclonic epilepsy (JME). JME is a form of idiopathic generalized epilepsy. It is yet unclear to what extent JME leads to abnormal network activation patterns. Our findings suggested that JME patients had altered multistability in selective functional networks and frequency bands in the frontoparietal cortices. |
| Exploitation Route | Computational modelling and simulations are critical analysis tools in contemporary neuroscience. Models at various levels of abstraction, or brain organisation attempt to capture different neuronal, or cognitive phenomena. In this project, we model first a voluntary decision process in the task, where two available options carry the same outcome reward probability. Trial-by-trial accumulation rates are modulated by single-trial EEG features. With hierarchical Bayesian parameter estimation, we show that the probability of reward was associated with changes in the speed of evidence accumulation. Also, we use pairwise Maximum Entropy Model (pMEM) to quantify irregularities in the MEG resting-state networks between juvenile myoclonic epilepsy (JME) patients and healthy controls. We show that JME group exhibited on average fewer local minima of the pMEM energy landscape than controls in the fronto-parietal network. Our results highlighted the pMEM as a descriptive, generative model for characterising atypical functional network properties in brain disorders. Next, we use hierarchical drift-diffusion model (HDDM) to study the integration of information between multiple information sources. We observe non-perfect integration in case of both congruent and incongruent evidence accumulation. Based on the HDDM parameters fit, we hypothesise about the neuronal implementation by extending a biologically plausible neuronal mass model of decision-making. Overall, this work paves the way for closer integration of theoretical models with behavioural and neuroimaging data. |
| Sectors | Healthcare,Pharmaceuticals and Medical Biotechnology |
| URL | https://www.mitpressjournals.org/doi/abs/10.1162/netn_a_00125,http://compneuro.uwaterloo.ca/files/publications/duggins.2020.pdf,https://link.springer.com/article/10.1007/s42113-020-00096-6 |