abstract: Understanding where our bodies are in space is imperative for motor control, particularly for actions such as goal-directed reaching. Multisensory integration is crucial for reducing uncertainty in arm position estimates. This dissertation examines time and frequency-domain correlates of visual-proprioceptive integration during an arm-position maintenance task. Neural recordings were obtained from two different cortical areas as non-human primates performed a center-out reaching task in a virtual reality environment. Following a reach, animals maintained the end-point position of their arm under unimodal (proprioception only) and bimodal (proprioception and vision) conditions. In both areas, time domain and multi-taper spectral analysis methods were used to quantify changes in the spiking, local field potential (LFP), and spike-field coherence during arm-position maintenance. In both areas, individual neurons were classified based on the spectrum of their spiking patterns. A large proportion of cells in the SPL that exhibited sensory condition-specific oscillatory spiking in the beta (13-30Hz) frequency band. Cells in the IPL typically had a more diverse mix of oscillatory and refractory spiking patterns during the task in response to changing sensory condition. Contrary to the assumptions made in many modelling studies, none of the cells exhibited Poisson-spiking statistics in SPL or IPL. Evoked LFPs in both areas exhibited greater effects of target location than visual condition, though the evoked responses in the preferred reach direction were generally suppressed in the bimodal condition relative to the unimodal condition. Significant effects of target location on evoked responses were observed during the movement period of the task well. In the frequency domain, LFP power in both cortical areas was enhanced in the beta band during the position estimation epoch of the task, indicating that LFP beta oscillations may be important for maintaining the ongoing state. This was particularly evident at the population level, with clear increase in alpha and beta power. Differences in spectral power between conditions also became apparent at the population level, with power during bimodal trials being suppressed relative to unimodal. The spike-field coherence showed confounding results in both the SPL and IPL, with no clear correlation between incidence of beta oscillations and significant beta coherence. / Dissertation/Thesis / Doctoral Dissertation Biomedical Engineering 2017

In everyday life we regularly update our expectations about the locations at which sensory events may occur, and about the motor responses that are appropriate in a given situation. The experiments in this thesis investigated the neural correlates of perceptual processes and motor preparation during human decision making, and the regions that causally contribute to decision making in the human brain. In Chapter 3, I used electroencephalography (EEG) to investigate whether alpha-band (~8-14 Hz) oscillations provide a graded index of participants’ preparatory attentional states. Time-frequency analysis revealed that manipulating spatial certainty regarding the location of an upcoming visual target led to parametric changes in the lateralization of preparatory occipito-parietal alpha oscillations, and to parametric modulation of parieto-central beta-band (~15-25 Hz) power typically associated with response preparation. In Chapter 4, I used EEG to investigate whether evolution of lateralization of sensorimotor alpha- and beta-band activity reflected participants’ evolving expectations about an upcoming motor response. Lateralization of activity in both frequency bands varied parametrically with the available evidence, suggesting such lateralized activity correlates with participants’ internal decision variables. Further analysis identified unique contributions to lateralized and non-lateralized oscillatory activity due to the prior evidence, evidence update, and surprise related to the observed information at each stage of the task. In Chapter 5, I extended the paradigm developed in Chapter 4 for use with online repetitive transcranial magnetic stimulation (TMS) and concurrent EEG recording. Delivery of TMS during decision making allowed investigation of the causal role played by a left hemisphere medial intraparietal region that is the putative human homologue of the macaque medial intraparietal cortex (MIP). MIP stimulation disrupted decision-making behaviour by biasing participants’ decisions against contralateral-to-stimulation (i.e., right-handed) responses. Comparison of the magnitude of TMS-induced changes in behaviour and beta-band activity demonstrated that the intraparietal cortex plays a causal role both in decision making and in the appearance of beta-band activity over the motor cortex. In Chapter 6, the broader consequences of the experimental work presented in this thesis are discussed, in addition to promising directions for future research.