Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 7 (10), e47992

Prediction of Muscle Activities From Electrocorticograms in Primary Motor Cortex of Primates

Affiliations

Prediction of Muscle Activities From Electrocorticograms in Primary Motor Cortex of Primates

Duk Shin et al. PLoS One.

Erratum in

  • PLoS One. 2014;9(3):e92653

Abstract

Electrocorticography (ECoG) has drawn attention as an effective recording approach for brain-machine interfaces (BMI). Previous studies have succeeded in classifying movement intention and predicting hand trajectories from ECoG. Despite such successes, however, there still remains considerable work for the realization of ECoG-based BMIs as neuroprosthetics. We developed a method to predict multiple muscle activities from ECoG measurements. We also verified that ECoG signals are effective for predicting muscle activities in time varying series when performing sequential movements. ECoG signals were band-pass filtered into separate sensorimotor rhythm bands, z-score normalized, and smoothed with a Gaussian filter. We used sparse linear regression to find the best fit between frequency bands of ECoG and electromyographic activity. The best average correlation coefficient and the normalized root-mean-square error were 0.92±0.06 and 0.06±0.10, respectively, in the flexor digitorum profundus finger muscle. The δ (1.5∼4Hz) and γ2 (50∼90Hz) bands contributed significantly more strongly than other frequency bands (P<0.001). These results demonstrate the feasibility of predicting muscle activity from ECoG signals in an online fashion.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1
Figure 1. Behavioral task and ECoG electrode locations.
A) Monkeys performed sequential right arm and hand movements, which consisted of reaching to a knob, grasping the knob with a lateral grip, pulling the knob closer, releasing the knob, and returning the hand to the home position, in a 3-D workspace. During the task, ECoG and EMG signals were recorded simultaneously. B) Schematic diagrams of ECoG electrode locations in left hemisphere. The planar-surface platinum electrode arrays were placed on the gyrus between the central sulcus (CS) and the arcuate sulcus (AS) in the primary motor area. The # indicates the location according to the column of ECoG electrodes.
Figure 2
Figure 2. Example of measured signals from monkey A during the tasks.
All signals are aligned to trial onset. At the top are frequency feature values of the ECoG signals. Frequency features were resorted in time and sensorimotor rhythm bands. Frequency features of each band are ordered by channel. Below the frequency features are the 12 EMG signals recorded from wire electrodes implanted into muscles of the right forelimb. The blue traces represent original muscle activities. The red traces represent muscle activities obtained by low pass filtering (cut-off frequency: 4 Hz). Below the EMG are grip force on the knob and logical signals, indicating presence of the monkey's hand on the home button or grasping the knob.
Figure 3
Figure 3. Representative example of predicted and recorded muscle activities.
Dotted lines are actual muscle activities from EMG signals measured with wire electrodes, and solid lines represent predicted muscle activities from ECoG signals of monkey A. The normalized root mean square error (nRMSE) and correlation coefficient (CC) are also shown above each panel.
Figure 4
Figure 4. CC and nRMSE distributions for each muscle of monkey A.
The height of each blue bar is equal to the CC density of the interval (0.05). The height of each red bar is equal to the nRMSE density of the interval (0.02). The total area of the histogram is equal to the number of trials used as validation data. Each dotted line with a number shows the median of nRMSE or CC for each muscle. For visualization, we substituted zeros for all negative CC values in validation.
Figure 5
Figure 5. Example of muscle activity prediction in a continuous time series from monkey B.
Dotted lines are actual muscle activities from EMG signals and solid lines are predicted muscle activities from ECoG signals over a 50 s time interval. Both lines were normalized to the ranges of actual muscle activities. The normalized root mean square error (nRMSE) and correlation coefficient (CC) are also shown.
Figure 6
Figure 6. Contribution of each frequency band for EMG prediction.
Each panel shows results of multiple comparisons among the frequency bands for each location level (A: monkey A, B: monkey B). Each bar represents the mean of the median weights of each frequency band. At the top of each graph are the sum of the total weights. The # indicates the location of the ECoG electrodes. Noteworthy significant differences between weight values of frequency bands are marked with * (p<0.05) and ** (p<0.001). Other significance comparisons are omitted for visualization purposes.
Figure 7
Figure 7. Simple main effect of electrode location contributing to EMG prediction.
Each panel shows results of multiple comparisons among the locations for each frequency band level (A: monkey A, B: monkey B). Each marker displays the mean of the median weights. Faded lines show non-significant frequency bands. Solid lines represent a significant difference between weight values (p<0.001) and the dotted lines represent no significance. The # indicates the location of the ECoG electrodes.

Similar articles

See all similar articles

Cited by 23 PubMed Central articles

See all "Cited by" articles

References

    1. Wolpaw JR, Mcfarland DJ, Neat GW, Forneris CA (1991) An Eeg-Based Brain-Computer Interface for Cursor Control. Electroen Clin Neuro 78: 252–259. - PubMed
    1. Pfurtscheller G, Neuper C (1994) Event-related synchronization of mu rhythm in the EEG over the cortical hand area in man. Neurosci Lett 174: 93–96. - PubMed
    1. Birbaumer N, Ghanayim N, Hinterberger T, Iversen I, Kotchoubey B, et al. (1999) A spelling device for the paralysed. Nature 398: 297–298. - PubMed
    1. Wolpaw JR, Birbaumer N, McFarland DJ, Pfurtscheller G, Vaughan TM (2002) Brain-computer interfaces for communication and control. Clin Neurophysiol 113: 767–791. - PubMed
    1. Khan YU, Sepulveda F (2010) Brain-computer interface for single-trial EEG classification for wrist movement imagery using spatial filtering in the gamma band. Iet Signal Processing 4: 510–517.

Publication types

Grant support

This study is supported by the Strategic Research Program for Brain Sciences from the MEXT of Japan. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Feedback