Limits...
The role of ECoG magnitude and phase in decoding position, velocity, and acceleration during continuous motor behavior.

Hammer J, Fischer J, Ruescher J, Schulze-Bonhage A, Aertsen A, Ball T - Front Neurosci (2013)

Bottom Line: The decoding accuracy based only on phase information was substantially (at least 2 fold) and significantly higher than that based only on magnitudes for position, velocity and acceleration.There was also no indication of additional magnitude-based mechanisms encoding movement information in the LFC range.Thus, our study contributes to elucidating the nature of the informative LFC of motor cortical population activity and may hence contribute to improve decoding strategies and BMI performance.

View Article: PubMed Central - PubMed

Affiliation: Bernstein Center Freiburg, University of Freiburg Freiburg, Germany.

ABSTRACT
In neuronal population signals, including the electroencephalogram (EEG) and electrocorticogram (ECoG), the low-frequency component (LFC) is particularly informative about motor behavior and can be used for decoding movement parameters for brain-machine interface (BMI) applications. An idea previously expressed, but as of yet not quantitatively tested, is that it is the LFC phase that is the main source of decodable information. To test this issue, we analyzed human ECoG recorded during a game-like, one-dimensional, continuous motor task with a novel decoding method suitable for unfolding magnitude and phase explicitly into a complex-valued, time-frequency signal representation, enabling quantification of the decodable information within the temporal, spatial and frequency domains and allowing disambiguation of the phase contribution from that of the spectral magnitude. The decoding accuracy based only on phase information was substantially (at least 2 fold) and significantly higher than that based only on magnitudes for position, velocity and acceleration. The frequency profile of movement-related information in the ECoG data matched well with the frequency profile expected when assuming a close time-domain correlate of movement velocity in the ECoG, e.g., a (noisy) "copy" of hand velocity. No such match was observed with the frequency profiles expected when assuming a copy of either hand position or acceleration. There was also no indication of additional magnitude-based mechanisms encoding movement information in the LFC range. Thus, our study contributes to elucidating the nature of the informative LFC of motor cortical population activity and may hence contribute to improve decoding strategies and BMI performance.

No MeSH data available.


Selection of optimal frequency ranges for velocity decoding. Representative results are shown from subject S1. (A) Search for low-pass filter cut-off frequency in time-domain ECoG LFC decoding. Mean correlation coefficient as a function of upper bound low-pass filter and a time offset τ between predictors (LFC ECoG feature vector) and decoded movement velocity. The highest correlation coefficient based on the LFC was found in the frequency range [0.1–1.5] Hz in all subjects. (B) Cumulative sum (bottom-up direction) of predicted velocity frequency components at different carrier frequencies f (FCf), decoded using time-frequency domain ECoG FDf features (DAlg3, Figure 3). Mean correlation coefficient between the cumulated predictions (i.e., FCf0, FCf0 + FCf1, FCf0 + FCf1 + FCf2, …) and actual velocity with different time offsets τ between predictors and movement velocity (as in A). Decoding accuracy saturated in all subjects after including the first 5 FCf.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3814578&req=5

Figure 5: Selection of optimal frequency ranges for velocity decoding. Representative results are shown from subject S1. (A) Search for low-pass filter cut-off frequency in time-domain ECoG LFC decoding. Mean correlation coefficient as a function of upper bound low-pass filter and a time offset τ between predictors (LFC ECoG feature vector) and decoded movement velocity. The highest correlation coefficient based on the LFC was found in the frequency range [0.1–1.5] Hz in all subjects. (B) Cumulative sum (bottom-up direction) of predicted velocity frequency components at different carrier frequencies f (FCf), decoded using time-frequency domain ECoG FDf features (DAlg3, Figure 3). Mean correlation coefficient between the cumulated predictions (i.e., FCf0, FCf0 + FCf1, FCf0 + FCf1 + FCf2, …) and actual velocity with different time offsets τ between predictors and movement velocity (as in A). Decoding accuracy saturated in all subjects after including the first 5 FCf.

Mentions: To set the optimal cut-off of the low-pass filter for the ECoG data in DAlg1, we conducted a search over this parameter (example for velocity in Figure 5A). Consistently across all subjects, we found a global maximum in the decoding accuracy (DA) at 1.5 Hz, which was then used to illustrate decoding and tuning analysis results.


The role of ECoG magnitude and phase in decoding position, velocity, and acceleration during continuous motor behavior.

Hammer J, Fischer J, Ruescher J, Schulze-Bonhage A, Aertsen A, Ball T - Front Neurosci (2013)

Selection of optimal frequency ranges for velocity decoding. Representative results are shown from subject S1. (A) Search for low-pass filter cut-off frequency in time-domain ECoG LFC decoding. Mean correlation coefficient as a function of upper bound low-pass filter and a time offset τ between predictors (LFC ECoG feature vector) and decoded movement velocity. The highest correlation coefficient based on the LFC was found in the frequency range [0.1–1.5] Hz in all subjects. (B) Cumulative sum (bottom-up direction) of predicted velocity frequency components at different carrier frequencies f (FCf), decoded using time-frequency domain ECoG FDf features (DAlg3, Figure 3). Mean correlation coefficient between the cumulated predictions (i.e., FCf0, FCf0 + FCf1, FCf0 + FCf1 + FCf2, …) and actual velocity with different time offsets τ between predictors and movement velocity (as in A). Decoding accuracy saturated in all subjects after including the first 5 FCf.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3814578&req=5

Figure 5: Selection of optimal frequency ranges for velocity decoding. Representative results are shown from subject S1. (A) Search for low-pass filter cut-off frequency in time-domain ECoG LFC decoding. Mean correlation coefficient as a function of upper bound low-pass filter and a time offset τ between predictors (LFC ECoG feature vector) and decoded movement velocity. The highest correlation coefficient based on the LFC was found in the frequency range [0.1–1.5] Hz in all subjects. (B) Cumulative sum (bottom-up direction) of predicted velocity frequency components at different carrier frequencies f (FCf), decoded using time-frequency domain ECoG FDf features (DAlg3, Figure 3). Mean correlation coefficient between the cumulated predictions (i.e., FCf0, FCf0 + FCf1, FCf0 + FCf1 + FCf2, …) and actual velocity with different time offsets τ between predictors and movement velocity (as in A). Decoding accuracy saturated in all subjects after including the first 5 FCf.
Mentions: To set the optimal cut-off of the low-pass filter for the ECoG data in DAlg1, we conducted a search over this parameter (example for velocity in Figure 5A). Consistently across all subjects, we found a global maximum in the decoding accuracy (DA) at 1.5 Hz, which was then used to illustrate decoding and tuning analysis results.

Bottom Line: The decoding accuracy based only on phase information was substantially (at least 2 fold) and significantly higher than that based only on magnitudes for position, velocity and acceleration.There was also no indication of additional magnitude-based mechanisms encoding movement information in the LFC range.Thus, our study contributes to elucidating the nature of the informative LFC of motor cortical population activity and may hence contribute to improve decoding strategies and BMI performance.

View Article: PubMed Central - PubMed

Affiliation: Bernstein Center Freiburg, University of Freiburg Freiburg, Germany.

ABSTRACT
In neuronal population signals, including the electroencephalogram (EEG) and electrocorticogram (ECoG), the low-frequency component (LFC) is particularly informative about motor behavior and can be used for decoding movement parameters for brain-machine interface (BMI) applications. An idea previously expressed, but as of yet not quantitatively tested, is that it is the LFC phase that is the main source of decodable information. To test this issue, we analyzed human ECoG recorded during a game-like, one-dimensional, continuous motor task with a novel decoding method suitable for unfolding magnitude and phase explicitly into a complex-valued, time-frequency signal representation, enabling quantification of the decodable information within the temporal, spatial and frequency domains and allowing disambiguation of the phase contribution from that of the spectral magnitude. The decoding accuracy based only on phase information was substantially (at least 2 fold) and significantly higher than that based only on magnitudes for position, velocity and acceleration. The frequency profile of movement-related information in the ECoG data matched well with the frequency profile expected when assuming a close time-domain correlate of movement velocity in the ECoG, e.g., a (noisy) "copy" of hand velocity. No such match was observed with the frequency profiles expected when assuming a copy of either hand position or acceleration. There was also no indication of additional magnitude-based mechanisms encoding movement information in the LFC range. Thus, our study contributes to elucidating the nature of the informative LFC of motor cortical population activity and may hence contribute to improve decoding strategies and BMI performance.

No MeSH data available.