Attendees at the fifth annual Center for Mind and Brain Idea Blitz heard exciting talks from five CMB faculty members.
Dr. Simona Ghetti kicked off the event by describing new neuroimaging work on infantile amnesia. The study uses a clever paradigm to examine if an increase in dentate gyrus and CA3 volume positively or negatively predicts memory performance. Toddlers are asleep during the scan, but the team is able to get functional measures in addition to the structural by playing a song the participants had previously encountered either forward or in reverse.
Dr. Tony Shahin spoke next challenging the traditional explanation of the McGurk effect. The McGurk effect is observed when participants view a speaker whose lip patterns do not match the audio signal (usually “ba” versus “pa”). This effect is usually attributed to fusion of the signal in auditory and visual cortex, but Dr. Shahin proposes a model where visual cortex directly affects the representation in auditory cortex.
Next, Dr. John Henderson described a new neuroimaging technique with an exciting name. FIRE (FIxation REgistered) fMRI uses co-registered eye-tracking and fMRI signals. Dr. Henderson will use this technique to examine prediction in language. While participants are reading a paragraph, the fMRI signal can be related to how well the fixated work fits expectations and how constrained (predictable) it is.
Dr. John Olichney gave a rapid fire presentation of his new R01, which will use ERPs to improve detection and treatment of Alzheimers Disease. Dr. Olichney explained that ERPs are sensitive, and thus may aid in early detection of Alzheimers Disease, provide a measure of treatment response, track disease progress, or detect potential mechanisms for how therapies work.
Finally, Dr. Steve Luck asked how we are able to complete tasks we had never done before. For example, if a friend asks us to “get the big blue bowl from the cabinet to the right of the microwave?” how are we able to complete this task with no training or feedback? Dr. Luck explained that while humans can complete these novel tasks easily, neural network models generally require lots of training and cannot rapidly switch between tasks. He plans to use a combinatorial task generator to make an enormous task space for undergraduates to complete in the lab (with no practice or feedback) so that he can propose biologically plausible neural network models to explain this behavior.
ucdneuroscience 