Psychology and Neuroscience Faculty Database
Psychology and Neuroscience
Arts & Sciences
Duke University

 HOME > Arts & Sciences > pn > Faculty    Search Help Login pdf version printable version 

Publications [#323342] of Miguel A. Nicolelis

search PubMed.

Journal Articles

  1. Yun, K; Lebedev, M; Nicolelis, MAL (2007). Prediction of motor timing using nonlinear analysis of local field potentials. Ifmbe Proceedings, 14(1), 1005-1008. [doi]
    (last updated on 2023/06/01)

    Abstract:
    Recently we have shown that temporal intervals preceding movement onset in a self-timed motor task can be accurately predicted from the ensembles activity of cortical neurons recorded in motor and premotor cortex. The aim of this study was to predict behavioral time from local field potentials (LFPs) recorded by the same electrodes. LFPs reflect mainly postsynaptic potentials. Since LFPs integrate activity of many neurons surrounding the electrode, they can contain information about motor preparation and execution, and the temporal intervals involved in the motor task. In particular, LFP oscillations that are modulated during movements have been reported in monkey motor cortex. In our experiment, the monkeys performed self paced movements of their hands. They had to withhold movements for at least 2.5s to receive a reward. This task was accompanied by modulations of the LFP frequency in the 10-20 Hz range. Nonlinear analysis methods – approximate entropy (ApEn), correlation dimension (CD), and largest Lyapunov exponent (LLE) – were used for quantify complexity (ApEn), dimensional dynamics (CD), and chaotic dynamics (LLE) in the LFP activity. ApEn and CD were modulated during movements. ApEn and CD were higher in task periods preceding movement onset. LLE also showed significant modulations: it was lower during the delay period of the task. Thus, nonlinear features of LFPs are correlated with motor timing. We suggest using nonlinear analyses of LFPs as a supplement of algorithms based on neuronal data in the design of brain-machine interfaces.


Duke University * Arts & Sciences * Faculty * Staff * Grad * Postdocs * Reload * Login