Investigating the function of alpha frequency oscillatory activity

Abstract

A fundamental challenge in modern neuroscience is to understand the role of synchronous oscillatory activity of groups of neurons in information processing. This thesis addressed the problem of how alpha frequency oscillatory activity might help control the flow of information from both the external world and from higher cognitive areas (responsible for inhibitory control, top-down and bottom-up information flow). A series of experiments investigated how alpha neuronal dynamics might aid/control cognition. In order to study the functional significance of alpha frequency oscillatory activity, the effects on performance in cognitive tasks of alpha activity directly elicited using photic stimulation were examined. Initially, we were interested in the role of alpha oscillations in information transfer across cortical areas, which was probed using a numerical Stroop task with every trial preceded by a flicker prime. The incongruent trials of the Stroop task introduce a conflict between competing responses which results in people being slower in responses to the task compared with congruent trials. That slower response has been related to increased communication between conflict processing fronto-parietal and early somatosensory regions. If alpha oscillations improve communication efficiency across the cortex it was predicted that inducing stronger alpha oscillations would affect the performance, (i.e. the Stroop cost would diminish). That hypothesis was tested in a series of three experiments. None of the manipulations (different frequencies, amplitudes induced and alpha phases where the Stroop task was initiated) showed that alpha oscillatory activity reduces the Stroop effect. However, the last task showed that people were faster when the task was preceded by an alpha frequency flicker prime, especially around 10Hz. The fourth experiment built on the well-established phenomena that when alpha activity is elicited in a particular hemisphere it attenuates processing of sensory information in that hemisphere, while the opposite hemisphere, is characterised by increased efficiency of information processing/flow. The study tested whether that could occur within a hemisphere by localised entrainment of part of the visual field. This hypothesis was tested by examining whether it could resolve differences in results previously published by Mathewson et al., (2012) and Spaak et al., (2014). In this study, a target circle was presented at time points after the offset of an alpha flicker prime, such that it was either in or out of phase with the prime. The target was displayed briefly, and then a masking ring appeared around the target location. There were two experimental conditions. First priming occurred at the central target location, and this was expected to inhibit perception at that location, (i.e. the target would be best detected at out of alpha phase time points). In contrast, in the second condition, the target surround area (e.g. the mask location) was stimulated, and this was expected to inhibit perception at that location, (i.e. the mask would be most effective in phase time points and so the target more easily detected). However, in both instances, target detection was best at in-phase time points and attenuated at out of phase time points, in line with Mathewson et al., (2012) results. This gives us some insight into the role of the alpha phase in allowing the external stimuli to be perceived/detected. The fifth experiment tested whether the level of spatial uncertainty of briefly presented target determines the alpha phase position for its best detection. This task used a similar masked circle paradigm as the fourth experiment, but the target could appear at one of two locations either side of fixation, which were both preceded by a flicker prime (either alpha frequency or randomly jittered) and followed by masking rings. The hypothesis was that the optimal alpha phase for target detection depends on whether people are pre-guided (by an arrow cue) to the target location or uncued (a higher level of spatial uncertainty). This hypothesis was again tested by examining whether it could resolve differences in results previously published by Mathewson et al., (2012) and Spaak et al., (2014). This experiment showed that the level of spatial uncertainty of briefly presented target determines the optimal alpha phase for its detection. Targets whose location was not pre-guided were the most likely to be detected when presented at time points out of phase with the entrained alpha prime; targets whose location was pre-guided by a brief arrow were the most likely to be detected when presented at time points in phase with the entrained alpha prime. The sixth experiment used EEG to investigate the neural dynamics underlying the behaviourally tested phenomenon in the previous experiment. Results showed that for targets with a high level of spatial uncertainty, the average alpha power peak was detected earlier in anterior electrodes compared with posterior electrodes, which is consistent with a greater reliance on alpha top-down dynamics. In contrast, for targets at a spatially cued location, the average alpha power peak was detected earlier at posterior electrodes, which suggests a greater reliance on bottom-up alpha neuronal dynamics. In summary, this thesis confirmed that mid-alpha phase determines the probability of detection of a briefly presented target. Also, it showed that optimal alpha phase for detecting briefly presented target would differ depending on the level of spatial uncertainty of that target. Targets at non-predictable locations are more likely to be detected at a trough in the phase of alpha activity whilst those at cued locations are most likely to be detected in-phase. Hence, perception depends not only on the internal neuronal alpha dynamics but also on the type of the visual percept. This difference may highlight the role of two different neuronal alpha sources which dominate in the different scenarios. When the target location is uncertain, top-down alpha dynamics dominate. However, when the target location is pre-guided, bottom-up alpha dynamics dominate.