Spatiotemporal emergence of movement plans in the posterior

Download Report

Transcript Spatiotemporal emergence of movement plans in the posterior

Spatiotemporal emergence of movement
plans in the posterior parietal cortex during
eye-hand coordination
Arpan Banerjee
NYU
Sloan-Swartz Meeting 2008
Eye hand coordination
Stimulus
onset
Time
Target
selection time
Reaction time
Plan
Working Hypotheses
Either, eye-hand coordination requires
shared target representations for saccades
and reaches which may lead to correlation in
target selection times across brain areas
Or, coordination is primarily mediated by
coupling between the movement plans after
the target is selected
Recordings of spiking and LFP activity
Spikes
Rate
(Hz)
Time (msec)
Fields
Voltage
Parietal reach region (PRR)
Lateral intra-parietal area (LIP)
Dean (In progress)
Lazzaro (In progress)
Overview
Primary Goal:
• Unified analysis of spikes and fields to detect target
selection times.
Methods
• Statistical modeling of the specific patterns (variates)
encoding movement directions.
• Decode the movement direction (and Target selection
time) from the pattern (variate) given new data via
likelihood ratio tests
Future direction: RT
Modelling of LFP activity
Neural
recording
area
Autoregressive modelling with external input
Decoding Input
• Obtain the a’s (AR coefficients) by Burg’s
algorithm
• Model order (p) selection using Akaike criterion
• Input decoded
• Amplitude and latency estimated trial by trial by
maximizing the posterior
Likelihood estimation
• Neural signal modeling for each direction
of movement
• Accumulation of logarithm of likelihood
ratio
is obtained
Target selection time estimation
• Accumulated log-likelihood ratios
Fields
104 msec
Future directions
• Application of Sequential probability ratio tests to
determine the target selection time from
accumulated ratios Wald 1940
• Time variation of A R coefficients.
• Conditional intensity process will be considered
for spike activity
• Extension to multivariate recordings
Summary
• Target selection time can be computed
simultaneously from fields and spikes in a
unified framework.
• Covariation of target selection times
across different brain areas and with
saccade and reach reaction times can be
used to understand the coupling between
eye and hand movements.
Acknowledgements
•
•
•
•
•
•
•
•
Dr. Bijan Pesaran
Stephanie Lazzaro
Dr. Heather Dean
Boris Revechkis
Sam Gershman
Eva Tsui
Adam Weiss
The Swartz Foundation
Thank you!