How is spontaneous brain activity analysed?

The concept:

Generally, it is difficult to study the brain ‘at rest’ with the approach generally pursued in science when external manipulation (independent variable) is used to obtain informative measurements (dependent variable) about the object of interest, because it may suspend the resting state. A related detailed review can be found in a Special Issue of Human Brain Mapping (Laufs, 2008) (see Error: Reference source not found). Concurrent

electroencephalography (EEG) and functional magnetic resonance imaging (fMRI) allow the simultaneous measurement of brain activity from two angles. One modality can be chosen to be interpreted as the independent variable and the other one as the dependent variable.

Spontaneous activity can be studied without external manipulation. Particularly, the EEG data can describe spontaneously occurring epileptic discharges and can be treated as the independent variable, forming a regressor to interrogate the fMRI data (dependent variable).

The inverse is equally possible (Laufs et al., 2003b; Morillon et al., 2010), and fusion analyses (Rosa et al., 2010) attempt to treat EEG and fMRI data equally (Daunizeau et al., 2007;

Daunizeau et al., 2010; Laufs et al., 2008; Luessi et al., 2011; Valdes-Sosa et al., 2009). Please refer to Figure 3 in section Error: Reference source not found, starting on p. Error: Reference source not found (Laufs, 2012b) for a formalized approach to the different types of analyses.

The physiology of the signals and their conceptual link

Blood oxygen level-dependent (BOLD) functional magnetic resonance imaging (fMRI) measures the haemodynamic correlate of neuronal activity, probably local field potentials (LFP) and – to a lesser degree – multi unit activity (Logothetis et al., 2001). Accordingly, BOLD signal decreases are related to neuronal activity decreases (Shmuel et al., 2006). The BOLD signal may reflect also sub-threshold activity, simultaneous inhibition with excitation and the result of modulating afferent input of remote neurons (Nunez and Silberstein, 2000).

Scalp electroencephalogram (EEG) measures neuronal activity in the form of postsynaptic excitatory and inhibitory potentials (EPSP and IPSP) of pyramidal cells perpendicular to the cortical surface. The overlap between EPSP and LFP remains unknown, and fMRI and EEG are measurements not of identical neuronal processes, providing valuable complementary information. Because of this, the relationship between simultaneously acquired EEG and fMRI data needs further consideration.

In practice: Still, as a best guess it is generally assumed in EEG-fMRI studies that interictal paroxysmal EEG discharges are associated with a haemodynamic response analogous to that commonly used for modelling in event-related studies in cognitive neuroscience, commonly called the canonical event-related haemodynamic response function (canonical HRF).

Analyses based on that assumption have produced significant response patterns that are generally concordant with prior electroclinical data. Assessing this formally, we used a more flexible model of the event-related response, a Fourier basis set, to investigate the presence of other responses in relation to individual IED in 30 experiments in patients with focal epilepsy. We found significant responses that had a noncanonical time course in 37% of cases, compared with 40% for the conventional, canonical HRF-based approach. In two cases, the Fourier analysis suggested activations where the conventional model did not. The

noncanonical activations were almost always remote from the presumed generator of

epileptiform activity. In the majority of cases with noncanonical responses, the noncanonical responses in single-voxel clusters were suggestive of artifacts. We did not find evidence for IED-related noncanonical HRFs arising from areas of pathology, suggesting that the BOLD response to IED is primarily canonical. Noncanonical responses may represent a number of phenomena, including artefacts and propagated epileptiform activity (Lemieux et al., 2008) .

When contrasting spontaneously occurring events against rest, the risk of false-positive activation is potentially higher than in studies with block or optimized event-related designs.

False positives – or negatives – due to image degradation can for example arise from subject motion which is usually worse in patients than in healthy subjects. Because discarding patient data is usually not an options, strategies need to be pursued optimising the yield of each experiment while minimizing the risk of false-positive activation. Afraim Salek-Haddadi developed a method, “scan nulling” (Salek-Haddadi et al., 2006) which models “jerky” head motion in addition to the established regression of scan-to-scan subject movement (Friston et al., 1996). We evaluated the efficacy of this approach by mapping the proportion of the brain for which F-tests across the additional regressors were significant. In 95% of cases, there was a significant effect of motion in 50% of the brain or greater; for the scan nulling effect, the proportion was 36%; this effect was predominantly in the neocortex. We concluded that the proposed approach is effective (Lemieux et al., 2007).

Another source of noise in resting state data is of physiological origin and due to cardiac pulsation and respiration and several methods have been proposed for their reduction (Birn et al., 2006; Birn et al., 2008; Glover et al., 2000; Laufs et al., 2007a) (see Error: Reference source not found). We also developed a method to reduce cardiac noise modelling it as an effect of no interest. Our model is based on an over-complete basis set covering a linear relationship between cardiac-related MR signal and the phase of the cardiac cycle or time after pulse (TAP). This method showed that, on average, 24.6 +/- 10.9% of grey matter voxels contained significant cardiac effects. 22.3 +/- 24.1% of those voxels exhibiting significantly interictal epileptic discharge-correlated BOLD signal also contained significant cardiac effects.

We quantified the improvement of the TAP model over the original model without cardiac effects, by evaluating changes in efficiency. Over voxels containing significant cardiac-related

signal efficiency was improved by 18.5 +/- 4.8%. Over the remaining voxels, no improvement was demonstrated (Liston et al., 2006b) (see Error: Reference source not found).