Olivia Gozel

Olivia Gozel

Faculty Mentor: John Maunsell, PhD
Department of Neurobiology, Harvard Medical School
Olivia Gozel

Project Summary: The neuronal mechanisms of attention take place without our awareness. Nevertheless, they play a critical role in determining what we experience. In the visual system, attention filters sensory stimuli by emphasizing important objects or locations and by reducing the impact of competing distractors.
If a location in the visual field is attended, then neurons whose receptive fields cover this space will have a higher spiking rate, lower variability of their activity, and decreased interneuronal correlations compared to those of an unattended location. However, less work has been done on the local field potentials (LFP) correlates of visual attention. It is of primary interest to fill this gap, as it is computationally less demanding to work with LFPs than with spiking activity because LFPs do not require spike sorting algorithms. Additionally, LFPs are more closely related to non-invasive techniques such as electroencephalograms which are more applicable to human patients.
To achieve a better understanding of visual attention, microelectrode arrays were bilaterally implanted into area V4 in the visual cortex of rhesus monkeys. These arrays were used to continuously record LFPs while the animal performed a task that controlled its visual attention. Using this approach, it is possible to connect brain activity with changes in how the animal directs its attention, and to relate these changes to behavioral performance. To this end, we are analyzing the LFP data in different ways. Power spectra will be computed to observe differences in specific frequency bands between attended and unattended conditions. Time-frequency domain analysis will also be performed to assess changes in the power of different frequencies over time. Additionally, receptive field maps for each channel will be used to assess the stimulus selectivity and stability of the LFPs. Finally the ultimate goal will be to use simultaneous recordings of LFP from multiple electrodes to predict the animal’s behavioral performance for individual stimulus presentations.

Fellowships

Fellowship Year