SPECTRAL ANALYSIS OF THE CEREBRAL ACTIVITY

DURING VOLUNTARY MODULATION OF MENTAL STATES

A High Resolution EEG Study

J. Toppi

1,2

, F. Babiloni

, F. Cincotti

, F. De Vico Fallani

2,3

, G. Vecchiato

2,3

, S. Salinari

D. Mattia

and L. Astolfi

1,2

Dept. of Computer Science and Systems, Univ. of Rome “Sapienza”, Rome, Italy

IRCCS “Fondazione Santa Lucia”, Rome, Italy

Dept. of Physiology and Pharmacology, Univ. of Rome “Sapienza”, Rome, Italy

Keywords: High Resolution EEG, Spectral Cortical Maps, Mental Imagery Task, Correction for Multiple Comparisons,

False Discovery Rate correction, Bonferroni correction.

Abstract: In the neuroscience field, the use of advanced techniques of EEG recording and analysis, led to look for

adequate methodology to prevent type I errors, which occur in computing thousands of univariate tests in

order to highlight the brain areas in which significant activity arises. In this paper we illustrate the capability

of tracking the brain activity during tasks consisting in tennis playing imagery and spatial navigation

imagery, by using advanced high resolution EEG methodology accompanied by the use of appropriate

statistical techniques that takes into account the risk of the Type I errors. Results showed that in the Spatial

Navigation condition the power spectra activity is significantly different from the rest in the bilateral

parietal areas and left motor area, while in the Tennis condition the cortical activity differs from the rest in

bilateral parietal areas and in the left sensory-motor cortex. These preliminary findings are in partial

accordance with previous hemodynamic studies.

1 INTRODUCTION

The rationale beyond this study relies on some

considerations about the use of adequate statistical

techniques in the framework of the

neuroelectromagnetic brain mapping.

With the use of advanced EEG/MEG recording

setup, involving a high number of sensors and

thousands of sources, the issue of the protection

against the Type I errors that could occur during the

execution of a high number of univariate statistical

tests has become of relevance.

Considering that in neuroscience, thousands of

univariate tests are performed to highlight the brain

areas in which significant activity arises, to seek for

adequate methodology to prevent this type I errors

has remarkable importance.

In this paper we illustrate the capability of

tracking the brain activity during some mental

imagery tasks (Owen et al., 2006), by using

advanced high resolution EEG methodology in the

time and frequency domains, accompanied by the

use of appropriate statistical techniques that takes

into account the risk of the Type I errors.

2 METHODS

2.1 Experimental Design

Five healthy volunteers took part in the experiment.

All the subjects were informed about the aim of the

EEG recording and signed an informed consent.

Subjects seated in front of a monitor. They executed

one of the three fixed imagery task (Play Tennis,

Relax or Imagine to visit the rooms of your house)

according to the position of a red target on the

screen (Figure 1).

The experiment was divided into 6 sessions of 18

trials each (6 for each task), with events randomly

ordered within each session. We set a task length of

15s and an inter-trial interval of 2s.

A 61-channel system was used to record EEG

potentials by means of an electrode cap. Sampling

rate was 200 Hz. EEG signals were then band pass

285

Toppi J., Babiloni F., Cincotti F., De Vico Fallani F., Vecchiato G., Salinari S., Mattia D. and Astolﬁ L..

SPECTRAL ANALYSIS OF THE CEREBRAL ACTIVITY DURING VOLUNTARY MODULATION OF MENTAL STATES - A High Resolution EEG Study.

DOI: 10.5220/0003168902850288

In Proceedings of the International Conference on Bio-inspired Systems and Signal Processing (BIOSIGNALS-2011), pages 285-288

ISBN: 978-989-8425-35-5

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

filtered (1-45 Hz) and eye movements were removed

from recordings utilizing Independent Component

Analysis (ICA). For the EEG analysis we considered

the interval [5:10] seconds in the middle of task

execution.

Figure 1: Stimulation window used for the experiment.

Subjects were asked to execute a particular task of mental

imagery according to the position of a red target on the

screen.

2.2 High Resolution EEG

High-resolution EEG technologies have been

developed to enhance the spatial information content

of EEG activity.

Accurate estimates of the cortical current density

could be obtained by using adequately detailed

geometrical reconstruction of the main

compartments lying between the cortical generator

sources and the EEG sensors. These estimates can be

obtained by solving a linear problem (Babiloni et al.,

2005), by means of a transfer matrix (lead field

matrix) that mimics the effects of the volume

conductor. In mathematical terms the relationship

between the modeled sources x, the lead field matrix

A, the EEG measurements b and the noise n can be

written as

Ax = b + n (1)

The solution of this linear system provides an

estimation of the dipole source configuration x that

generates the measured EEG potential distribution b.

The system includes also the measurement noise

n, assumed to be normally distributed. A is the lead

field or the forward transmission matrix, whose j-th

column describes the potential distribution generated

on the scalp electrodes by the j-th unitary dipole.

The current density solution vector ξ was

obtained as:

(

)

minarg

A xbx

λξ

+−=

(2)

where M, N are the matrices associated to the

metrics of the data and of the source space,

respectively, λ is the regularization parameter and

|| x ||

represents the M norm of the vector x. The

solution of Eq. (2) is given by the inverse operator

Gbξ

(3)

(

)

111

−

−−−

′′

= MAANANG

(4)

An optimal regularization of this linear system was

obtained by the L-curve approach.

Using the relations described above, an estimate

of the signed magnitude of the dipolar moment for

each one of the 5,000 cortical dipoles was obtained

for each time point.

2.3 Spectral Cortical Activity

From the cortical waveforms, we estimated the

spectral activity, during the considered task time

interval, for each one of the 5 thousands dipoles of

the cortical model used.

T-test values obtained from comparisons

between Tennis-Rest and Navigation-Rest were then

mapped on the cortical model in different frequency

bands, defined according to Individual Alpha

Frequency (IAF) to take into account inter-

individual differences in localization of alpha band.

The IAF, defined as the individual frequency peak

within the alpha band, was determined from the Fast

Fourier Transform spectra over posterior leads

(parietal, parieto-occipital, and occipital).

Individually defined bands considered were:

Theta (IAF-6 / IAF-2), Alpha (IAF-2 / IAF+2), Beta

(IAF+2 / IAF+14) and Gamma (IAF+15 / IAF+30).

The uncorrected Student’s test and the appropriate

techniques of the False Discovery Rate (FDR) (Yoav

and Yekutieli, 2001) and the Bonferroni correction

for multiple comparisons were applied to the

evaluation of the power spectral maps estimated

from the data.

3 RESULTS

By following the procedure and the methods

illustrated above, we obtained statistical scalp and

cortical maps for each frequency band of interest

and experimental condition.

In the following figures, T-test values,

BIOSIGNALS 2011 - International Conference on Bio-inspired Systems and Signal Processing

286

Figure 2: Statistically significant spectral maps in the Alpha band, depicted on the average brain model used in the analysis.

The color scale codes for the value of the t-test in that pixel for the comparison Tennis versus Rest, in three different

cases: a) no correction; b) False Discovery Rate; c) Bonferroni.

corresponding to a statistical significance of 0.05,

were mapped on scalp and cortical average model in

three different cases: no correction for multiple

comparisons (panel a), FDR correction (panel b) and

Bonferroni correction (panel c). The head is seen

from above, with the nose pointing down. For each

comparison, only frequency bands with significant

statistical activations are presented.

In Fig. 2 we show results in a representative

subject, describing the statistically significant

spectral maps in the Alpha band for the contrast

Tennis task (T) versus Rest condition (R).

Scalp maps, obtained without applying any

corrections for multiple comparisons (Figure 2a)

reveal a decrease of EEG activity in Tennis task on

electrodes located in the posterior areas, in the

central part of right side of the head and in the

frontal region. Cortical maps lead to a better

localization of deactivated areas, in particular in the

bilateral parietal lobes and left motor cortex.

In both scalp and cortical maps the corrections

for multiple comparisons, needed to prevent Type I

errors, brings to a decrease of the number of

activated pixels on the cortex model. In particular,

low reduction in the size of activated areas is shown

for FDR correction (Figure 2b), while no significant

activations survive using Bonferroni methods

(Figure 2c).

In Fig. 3 we show results in a representative subject,

describing the statistically significant spectral maps

in the Alpha band for the contrast Navigation task

(N) versus Rest condition (R). Significant

activations on scalp model, not corrected for

multiple comparisons, result in left parietal and

central areas (Figure 3a). Cortical maps lead to a

better localization of deactivated area, in particular

in the bilateral parietal and left motor area. Low

reduction in the size of activated areas both on scalp

and cortical model, is shown for FDR correction

(Figure 3b). No significant activations were found

using Bonferroni correction for multiple

comparisons (Figure 3c).

4 CONCLUSIONS

Thanks to the high resolution EEG techniques and

the appropriate use of statistical methods, we tracked

the subject’s brain activity during different

imagination tasks. Analyzing the effect of the two

methods of correction for multiple comparisons on

SPECTRAL ANALYSIS OF THE CEREBRAL ACTIVITY DURING VOLUNTARY MODULATION OF MENTAL

STATES - A High Resolution EEG Study

287

Figure 3: Statistically significant spectral maps in the Alpha band, depicted on the average brain model used in the analysis.

The color scale codes for the value of the t-test in that pixel for the comparison Navigation versus Rest, in three different

cases: a) no correction; b) False Discovery Rate; c) Bonferroni.

the estimated areas, the technique of False

Discovery Rate results as a good compromise to

prevent both type I and type II errors. In fact in the

Bonferroni method, the increased incidence of false

negative led to non-physiological results.

As a whole, these results suggest that the high

resolution EEG spectral mapping opens a way to

address the analysis of brain imaginative functions,

allowing to discriminate between different mental

states without the limitations of the use of an fMRI

scanner.

ACKNOWLEDGEMENTS

This work was supported by the European ICT

Program FP7-ICT-2009-4 Grant Agreement 247919

DECODER.

REFERENCES

Babiloni F. et al., 2005. Estimation of the cortical

functional connectivity with the multimodal

integration of high resolution EEG and fMRI data by

Directed Transfer Function. Neuroimage.

Owen A. M. et al, 2006. Detecting Awareness in the

Vegetative State. Science.

Yoav B. and Yekutieli D., 2001. The control of the false

discovery rate in multiple testing under dependency.

The Annals of Statistics.

BIOSIGNALS 2011 - International Conference on Bio-inspired Systems and Signal Processing

288