Supervisors: Arash Sahraie, Thomas Otto, Mary-Joan MacLeod
Project Description:
This project investigate the characteristics of mechanisms and their neuronal underpinning for sensory integration, leading to conscious experience. In particular we are interested in integration between visual and auditory systems leading to awareness of events.
Signal Detection Theory has provided a solid ground to investigate stimulus detection across senses. Given a sensory system (e.g., audition or vision), we can quantify the level of signal that is needed (the light intensity or the sound amplitude) to be distinguished from the background and therefore to be detected. The detection capacity is then expressed as a probability function of the stimulus strength in that sense. It has been known for nearly a century that if two signals (e.g., both light flash and sound) are combined/coincidental, then the overall detection probability is different than that expected from each sense separately. This improved performance is termed redundancy gain.
A number of computational modelling approaches have been used to understand the mechanisms underlying redundancy gains (Otto et al 2013), including for example race and co-activation models. These approaches are extensions of the (unisensory) evidence accumulation framework to a scenario in which two sensory stimuli provide evidence for a perceptual decision. They use reaction time data and error rates to distil the stages of processing such as information uptake, bias, decision criteria or non-decisional components that are affected in decision making based on sensory processing. Interestingly, all the above are reliant on detection capacity and not conscious awareness.
An important discovery has been the dissociation of detection and awareness (Weiskrantz 2009). This is particularly relevant to cases of brain injury when a brain injured participant may be significantly above chance in detecting a signal whilst has no conscious experience of a stimulus (labelled blindsight, or deaf-hearing). Repeated exposure to sensory stimuli can lead to increased sensitivity and recovery of function. There has been one study showing enhanced performance after systematic stimulation using supra-threshold multi-sensory stimuli (Bolognini et al 2005). If the findings are confirmed this can open up a new frontier in developing technologies for rehabilitation after brain injury. However, both the underlying mechanisms and the extent of the influence of multi-sensory input on neuronal plasticity is not known.
The proposed project is a collaborative effort combining expertise from psychophysics, computational modelling and clinical Neuroscience of the three supervisors. The project will examine audio-visual interaction in the context of reported conscious awareness. Using psychophysical investigation, the parameters for such interactions will be mapped out. We will develop and apply computational models in relation to reported awareness to find out which aspects of processing could be affected by multi-sensory input. In addition, we intend to recruit participants with brain injury (stroke survivors) with lesions in either primary visual or auditory cortical areas in order to examine the model predictions. These basic investigations on neuronal encoding of sensory processes are crucial in forming a solid foundation for development of future rehabilitation techniques.
The successful candidate will have the opportunity to train on psychophysical and computational methods. By the end of the PhD, they will be competent in MatLab programming, experimental design and advance analysis techniques such as hierarchical linear models. They will participate fully in EastBio organised training events, including induction, foundation training, professional and advance training to build their skills set.
References:
Weiskrantz L. (2009) Blindsight, Oxford University Press.
Bolognini et al (2005) Brain 128 (12), 2830-2842.
Otto et al (2013) Journal of Neuroscience 33(17) 7463-74.
If you want to apply for this project, please go to this link.
