Limits...
Inferring functional brain states using temporal evolution of regularized classifiers.

Zhdanov A, Hendler T, Ungerleider L, Intrator N - Comput Intell Neurosci (2007)

Bottom Line: We approach the problem from a machine learning perspective, by constructing a classifier from a set of labeled signal examples.We propose a framework that focuses on temporal evolution of regularized classifiers, with cross-validation for optimal regularization parameter at each time frame.We demonstrate the inference obtained by this method on MEG data recorded from 10 subjects in a simple visual classification experiment, and provide comparison to the classical nonregularized approach.

View Article: PubMed Central - PubMed

Affiliation: Functional Brain Imaging Unit, Tel Aviv Sourasky Medical Center, 6 Weizmann Street, Tel Aviv 64239, Israel. zhdanova@post.tau.ac.il

ABSTRACT
We present a framework for inferring functional brain state from electrophysiological (MEG or EEG) brain signals. Our approach is adapted to the needs of functional brain imaging rather than EEG-based brain-computer interface (BCI). This choice leads to a different set of requirements, in particular to the demand for more robust inference methods and more sophisticated model validation techniques. We approach the problem from a machine learning perspective, by constructing a classifier from a set of labeled signal examples. We propose a framework that focuses on temporal evolution of regularized classifiers, with cross-validation for optimal regularization parameter at each time frame. We demonstrate the inference obtained by this method on MEG data recorded from 10 subjects in a simple visual classification experiment, and provide comparison to the classical nonregularized approach.

No MeSH data available.


(a) Temporal stabilityof the best separating timeslice as a function of regularization parameter forsubject JMB. The upper plot shows the accuracy of the classifier as a functionof timeslice and regularization parameter. The accuracy is denoted by the coloraccording to the colorbar above the plot. Timeslice yielding maximum accuracyfor each value of the regularization parameter is marked by a black dot. Thelower part of the plot shows the best (over all timeslices) error plottedagainst the regularization parameter using the same timescale as the upper part.(b) Same as (a) but for subject MKN.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2266829&req=5

fig6: (a) Temporal stabilityof the best separating timeslice as a function of regularization parameter forsubject JMB. The upper plot shows the accuracy of the classifier as a functionof timeslice and regularization parameter. The accuracy is denoted by the coloraccording to the colorbar above the plot. Timeslice yielding maximum accuracyfor each value of the regularization parameter is marked by a black dot. Thelower part of the plot shows the best (over all timeslices) error plottedagainst the regularization parameter using the same timescale as the upper part.(b) Same as (a) but for subject MKN.

Mentions: Another item of particular interest is the temporalstructure of the signal and its relation to the regularization parameter. Wediscovered that the stability of the best separating timeslice as a function ofregularization and classifier performance as a function of regularization areclosely related. The temporal location of the best separating timeslice tendsto be more stable for the λ values thatyield lower classification error (see Figure 6).


Inferring functional brain states using temporal evolution of regularized classifiers.

Zhdanov A, Hendler T, Ungerleider L, Intrator N - Comput Intell Neurosci (2007)

(a) Temporal stabilityof the best separating timeslice as a function of regularization parameter forsubject JMB. The upper plot shows the accuracy of the classifier as a functionof timeslice and regularization parameter. The accuracy is denoted by the coloraccording to the colorbar above the plot. Timeslice yielding maximum accuracyfor each value of the regularization parameter is marked by a black dot. Thelower part of the plot shows the best (over all timeslices) error plottedagainst the regularization parameter using the same timescale as the upper part.(b) Same as (a) but for subject MKN.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2266829&req=5

fig6: (a) Temporal stabilityof the best separating timeslice as a function of regularization parameter forsubject JMB. The upper plot shows the accuracy of the classifier as a functionof timeslice and regularization parameter. The accuracy is denoted by the coloraccording to the colorbar above the plot. Timeslice yielding maximum accuracyfor each value of the regularization parameter is marked by a black dot. Thelower part of the plot shows the best (over all timeslices) error plottedagainst the regularization parameter using the same timescale as the upper part.(b) Same as (a) but for subject MKN.
Mentions: Another item of particular interest is the temporalstructure of the signal and its relation to the regularization parameter. Wediscovered that the stability of the best separating timeslice as a function ofregularization and classifier performance as a function of regularization areclosely related. The temporal location of the best separating timeslice tendsto be more stable for the λ values thatyield lower classification error (see Figure 6).

Bottom Line: We approach the problem from a machine learning perspective, by constructing a classifier from a set of labeled signal examples.We propose a framework that focuses on temporal evolution of regularized classifiers, with cross-validation for optimal regularization parameter at each time frame.We demonstrate the inference obtained by this method on MEG data recorded from 10 subjects in a simple visual classification experiment, and provide comparison to the classical nonregularized approach.

View Article: PubMed Central - PubMed

Affiliation: Functional Brain Imaging Unit, Tel Aviv Sourasky Medical Center, 6 Weizmann Street, Tel Aviv 64239, Israel. zhdanova@post.tau.ac.il

ABSTRACT
We present a framework for inferring functional brain state from electrophysiological (MEG or EEG) brain signals. Our approach is adapted to the needs of functional brain imaging rather than EEG-based brain-computer interface (BCI). This choice leads to a different set of requirements, in particular to the demand for more robust inference methods and more sophisticated model validation techniques. We approach the problem from a machine learning perspective, by constructing a classifier from a set of labeled signal examples. We propose a framework that focuses on temporal evolution of regularized classifiers, with cross-validation for optimal regularization parameter at each time frame. We demonstrate the inference obtained by this method on MEG data recorded from 10 subjects in a simple visual classification experiment, and provide comparison to the classical nonregularized approach.

No MeSH data available.