Single-trial coupling of the gamma-band response and the corresponding BOLD signal
Introduction
Functional magnetic resonance imaging (fMRI) based on hemodynamic changes as assessed by the BOLD (blood oxygenation level dependent) signal is widely used today to indirectly investigate the neuronal activity of the brain. While the precise relationship between neuronal activity and the BOLD signal is not yet fully understood, several lines of evidence support an especially close coupling between high frequency oscillations in the gamma-band frequency range (> 30 Hz) and hemodynamic changes that can be assessed using the BOLD signal. The BOLD signal was shown to be most closely related to local field potentials in the gamma-band range using invasive measurements in animals (Logothetis et al., 2001). Corresponding evidence was also provided by studies in men demonstrating a closer relationship of the BOLD signal and gamma-band oscillations than between the BOLD signal and event-related-potentials (ERP) (Foucher et al., 2003). Accordingly, theoretical models were developed predicting that a shift in the frequency profile of the EEG to high frequency components should be correlated with an increase in the BOLD signal (Kilner et al., 2005).
Neuronal oscillations in the gamma-band frequency range have attracted much interest because they were suggested to play an important role in the linking of neurons into cell assemblies that code information in the brain (Singer, 1999). Experimental data obtained both in animals and in humans suggest that gamma-band oscillations are involved in perception and cognition (Herrmann et al., 2004a, Ribary, 2005). For example, several authors have described the early evoked gamma-band response (GBR) to be involved in attention (Tiitinen et al., 1993, Debener et al., 2003). Furthermore, an effect of task difficulty on the GBR amplitude was described (Mulert et al., 2007).
It is an important question, which neuronal substrates are responsible for the gamma-band oscillations measured on the scalp. The auditory GBR is at least partially generated in the auditory cortex (Pantev et al., 1991, Schadow et al., 2009) as a result of thalamo-cortical interactions (Ribary et al., 1991). However, since in cognitive tasks such as those mentioned above, it is well known that additional brain regions are involved, e.g. in the frontal lobe, one might wonder whether some of these regions might contribute as electrical generators to the GBR measured on the scalp or being functionally coupled by gamma-band oscillations. Using EEG source analysis, generators of the GBR in both the auditory cortex and the anterior cingulate cortex (ACC) have been described in more difficult conditions of a cognitive task (Mulert et al., 2007). However, EEG-based source imaging suffers from a limited spatial resolution in comparison to imaging techniques such as fMRI. In addition, it is difficult to investigate the role of sub-cortical structures such as the thalamus using scalp-EEG, although the thalamus is assumed to play a major role in the generation of brain rhythms (Steriade, 2006).
Recently, the integration of EEG and simultaneous fMRI was successfully introduced in order to combine the high temporal resolution of the EEG and the high spatial resolution of the fMRI (Ives et al., 1993, Lemieux et al., 1999, Goldman et al., 2000, Mulert et al., 2002b, Moosmann et al., 2003; for a review, see Ritter and Villringer, 2006, Mulert et al., 2008a). In particular, using single-trial coupling of EEG and fMRI, specific BOLD correlates of distinct neurophysiological components were demonstrated including the CNV (Nagai et al., 2004), N100 (Mulert et al., 2008b), P300 (Eichele et al., 2005, Benar et al., 2007) and ERN (Debener et al., 2005). Accordingly, it was the aim of the present study to use single-trial variations of the evoked gamma-band response to predict the GBR-related BOLD signal. We hypothesized to find GBR-related BOLD changes in brain regions that were described earlier using EEG source imaging during a difficult cognitive task (auditory cortex, ACC). In addition, we assumed to find GBR-related BOLD changes in the thalamus.
Section snippets
Subjects
Ten healthy subjects (4 men, 6 women, mean age = 22.8 years, range = 19–30 years, 9 right-hander, 1 left-hander) with no history of neurological and psychiatric disturbances or reduced hearing were recruited from an academic environment and participated in the study after giving their written informed consent. The research protocol was approved by the Ethics Committee for Human Experiments at the Ludwig-Maximilian University, Munich, Germany.
Paradigm
We used three different difficulty levels of an auditory
Reaction times
There was a significant condition effect for reaction times [F(2,18) = 34.7, p < 0.001] with longer reaction times in the more difficult runs. In the post hoc t-test analysis, reaction times were longer in DC than in IC [t(9) = 5.5, p < 0.001] and in EC [t(9) = 7.7, p < 0.001]. In addition, reaction times were longer in IC than in EC [t(9) = 3.75, p = 0.005], see Table 1.
Error rates
Error rates were very low in EC (0.5 %), higher in IC (3.4 %) and highest in DC (9.1 %). A significant condition effect was present [F
Discussion
This simultaneous EEG-fMRI study was intended to investigate the functional neuroanatomy of the auditory gamma-band response (GBR) using single-trial coupling of EEG and fMRI. We assumed both the auditory cortex and the thalamus to be involved in the generation of the GBR and aimed to clarify earlier reports of an involvement of the ACC. In accordance to our hypothesis, we found GBR-specific fMRI activations in the auditory cortex and the thalamus. In addition, we could also demonstrate
Acknowledgment
Part of this work was prepared in the context of Philip Hepp's dissertation at the Faculty of Medicine, Ludwig-Maximilians-University, Munich.
References (58)
- et al.
Suppression of transient 40-Hz auditory response by haloperidol suggests modulation of human selective attention by dopamine D2 receptors
Neurosci. Lett.
(2000) - et al.
Identification of EEG events in the MR scanner: the problem of pulse artifact and a method for its subtraction
Neuroimage
(1998) - et al.
Conflict monitoring and anterior cingulate cortex: an update
Trends Cogn. Sci.
(2004) - et al.
Analysis of fMRI time-series revisited
Neuroimage
(1995) - et al.
Reduced oscillatory gamma-band responses in unmedicated schizophrenic patients indicate impaired frontal network processing
Clin. Neurophysiol.
(2004) - et al.
Acquiring simultaneous EEG and functional MRI
Clin. Neurophysiol.
(2000) - et al.
Correlating the alpha rhythm to BOLD using simultaneous EEG/fMRI: inter-subject variability
Neuroimage
(2006) - et al.
Gamma responses and ERPs in a visual classification task
Clin. Neurophysiol.
(1999) - et al.
Cognitive functions of gamma-band activity: memory match and utilization
Trends Cogn. Sci.
(2004) - et al.
Source localization (LORETA) of the error-related-negativity (ERN/Ne) and positivity (Pe)
Brain Res. Cogn. Brain Res.
(2004)