Correlates of alpha rhythm in functional magnetic resonance imaging and near infrared spectroscopy

Abstract

We used simultaneous electroencephalogram-functional magnetic resonance imaging (EEG-fMRI) and EEG-near infrared spectroscopy (NIRS) to investigate whether changes of the posterior EEG alpha rhythm are correlated with changes in local cerebral blood oxygenation. Cross-correlation analysis of slowly fluctuating, spontaneous rhythms in the EEG and the fMRI signal revealed an inverse relationship between alpha activity and the fMRI-blood oxygen level dependent signal in the occipital cortex. The NIRS-EEG measurements demonstrated a positive cross-correlation in occipital cortex between alpha activity and concentration changes of deoxygenated hemoglobin, which peaked at a relative shift of about 8 s. Our data suggest that alpha activity in the occipital cortex is associated with metabolic deactivation. Mapping of spontaneously synchronizing distributed neuronal networks is thus shown to be feasible.

Introduction

Functional neuroimaging studies typically are designed to detect brain activity that occurs transiently after application of external or internal stimuli. However, several aspects of brain function do not follow such a “stimulus on/off” pattern. In particular, the electroencephalogram (EEG) during ongoing brain activity is dominated by spontaneously occurring, spatially distributed, oscillatory rhythms in characteristic frequency bands. These rhythms are presumed to correspond to the synchronized synaptic activity of large numbers of neurons. While there is still uncertainty about the functional role of these rhythms, one interesting hypothesis, the “binding phenomenon” (Bushara et al., 2003), supposes that the synchronization of distributed neuronal networks functionally integrates different brain structures. However, “visualization” of the brain regions involved is difficult based on EEG measurements alone. Whether high-resolution methods based on metabolic/vascular signals are capable of identifying the brain areas involved depends on whether a variation in the spectral power of these rhythms is accompanied by changes in the metabolism or blood flow.

In this study, we addressed this issue for the prototypical human EEG rhythm: the posterior alpha rhythm, which oscillates in the 8- to 12-Hz frequency range. It is the dominant rhythm in the human scalp EEG of relaxed but alert adults when the eyes are closed, disappearing when the eyes are open. First described by Hans Berger (Berger, 1929), the alpha rhythm is regarded as representing a resting or “idling” state of the underlying cortex (Adrian and Matthews, 1934).

Using a combined EEG-fMRI approach, two previous studies have provided the first data on blood oxygen level dependent (BOLD) correlates of alpha rhythm in an eye opening/closing task (Patel et al., 1997) and during spontaneous variations of the alpha rhythm (Goldman et al., 2002). Further elaborating on this approach, we took advantage of recent advances in simultaneous EEG-fMRI acquisition Lemieux et al., 1997, Sijbers et al., 1999, Allen et al., 2000, Hoffmann et al., 2000, Bonmassar et al., 2001b, Thees et al., 2003. To investigate the relationship between the alpha rhythm and potentially associated changes in a metabolic/vascular parameter, we employed a T2*-weighted MR pulse sequence known to produce a signal partially dependent on the concentration of deoxygenated hemoglobin ([deoxy-Hb]) BOLD signal (Ogawa et al., 1990). It should be emphasized, however, that the signal obtained by these MR sequences can be influenced by numerous other physiological factors. Thus, the well-established relationship between an increased MRI signal in T2*-weighted pulse sequences and the underlying decrease in [deoxy-Hb] during a typical “brain stimulation study” Kwong et al., 1992, Ogawa et al., 1992, Frahm et al., 1992, Bandettini et al., 1992 cannot be simply transferred to a study in which the alpha rhythm is “correlated” with the MRI signal. In order to test whether potential alpha-rhythm-associated MR signals truly reflect changes in [deoxy-Hb] we performed additional near infrared spectroscopy (NIRS) measurements, which permitted us to determine cerebral [deoxy-Hb] without interfering with the simultaneously acquired EEG signal (Obrig et al., 2002).

Specifically, we addressed the following questions:

• Is simultaneous recording of the alpha rhythm and fMRI-BOLD signal feasible?

And if so:

• Are there signal changes in T2*-weighted MRI pulse sequences that are reproducibly associated with changes in the alpha rhythm?

And if so:

• Are these fMRI signal changes explainable by changes in deoxy-Hb concentration?