VIGALL 2.0: Analyzing different functional brain states and their regulation during resting states

VIGALL 2.0 – a BrainVision Analyzer-compatible tool developed by the University of Leipzig for analysis of different functional brain states (vigilance stages) and their regulation during resting states

Vigilance (or brain arousal) strongly influences performance and neurophysiological reactions to stimuli and tasks. Dysregulation of vigilance is an important element in the pathophysiology of neuropsychiatric diseases. VIGALL 2.0 provides a tool to assess this basic aspect of brain function.

The human brain is a highly integrated network which is permanently interacting with itself. This network has different global functional states corresponding to different modes of function (similar to the different gears of a car). These can be separated based on the temporo-spatial pattern of scalp-recorded EEG and EOG activity. Such global functional brain states (levels of brain-arousal called “vigilance stages”) can be separated not only during sleep (e.g. slow wave sleep, REM) but also during the transition from active wakefulness to drowsiness and sleep onset.

The fundamental importance of the level and the precise and rapid adaptation and regulation of vigilance to the environment is obvious. Vigilance levels not only differ profoundly concerning EEG activity, but also determine our behavioral and neurophysiological response to stimuli and tasks (Bekhtereva et al., 2014; Minkwitz et al., 2011/2012). Vigilance regulation, i.e. the ability to maintain or alternate between vigilance stages, determines whether we can keep up our alertness and focused attention or whether we are able to “come down”, to relax and to fall asleep under defined conditions, which has importance e.g. for fMRI, PET or SPECT recordings.

However, research on this central neurophysiological mechanism and its role in cognitive functioning as well as its pathophysiologic relevance in different neuropsychiatric disorders has been hampered by the lack of valid and time economic classification tools. Objective tests for the regulation of vigilance like the Multiple Sleep Latency Test or the Maintenance of Wakefulness Test (Littner et al., 2005) are time-consuming and fail to provide information about the transitions between different vigilance stages during wakefulness.

Changes of EEG activity during the transition from active wakefulness to sleep onset are well described (Bente, 1964; Roth, 1961) and have been shown in many studies (Benca et al., 1999; Cantero et al., 2002; Corsi-Cabrera et al., 2000; De Gennaro et al., 2001a/b/2004; De Gennaro and Ferrara, 2003; Kaida et al., 2006; Marzano et al., 2007; Strijkstra et al., 2003; Tsuno et al., 2002). The following vigilance stages can be observed during the transition from high alertness to relaxed wakefulness to drowsiness and finally sleep onset (see figure 1):

• Stage 0: desynchronized non-alpha EEG in the absence of slow horizontal eye movements; found during an activated state (e.g. mental effort)
• Stage A1: with dominant alpha activity corresponding to relaxed wakefulness
• Stage A2: with spreading of alpha activity from occipital to more anterior cortices
• Stage A3: with pronounced frontal alpha activity; beginning of drowsiness
• Stage B1: non-alpha EEG with low amplitude (similar spectral composition to stage 0) but with presence of slow horizontal eye movements; drowsiness
• Stage B2/3: non-alpha EEG with predominant theta/delta activity, occasional occurrence of vertex waves; drowsiness and transition to sleep onset
• Stage C: commencing with occurrence of sleep spindles or K-complexes; sleep onset

Figure 1: Examples of the distinct vigilance stages that can be separated by the VIGALL algorithm, according to EEG and HEOG characteristics (see text for details).

The regulation of vigilance shows considerable inter-individual differences (see figure 2 for a schematic illustration). During a twenty-minute eyes-closed resting condition, most subjects show progressive declines to lower vigilance stages (adaptive vigilance regulation). However, while some subjects exhibit rapid declines within only a few seconds (unstable vigilance regulation), others continuously remain in stages of high vigilance (hyperstable vigilance regulation). This trait is modulated by many individual and environmental factors such as sleep deficits, vigilance enhancing substances (e.g. caffeine, nicotine), effort, motivation, and disease related factors. Therefore, care must be taken, to record EEG under comparable conditions and subjects are not to be distracted or awoken during the recording.

Figure 2: Schematic Illustration of the three main patterns of vigilance regulation and their association with psychiatric diseases. Under normal circumstances subjects show a gradually declining vigilance level when recorded under prolonges resting conditions without external stimulation (adaptive vigilance regulation). In some cases rapid drops to lower vigilance levels shortly after the beginning of the recording are seen (instable vigilance regulation) or, on the contrary, the decline in vigilance is delayed or even absent (hyperstable vigilance regulation).

Vigilance regulation is a promising diagnostic and prognostic biomarker for affective disorders, ADHD or fatigue (Geissler et al., 2014; Hegerl et al., 2013; Schönknecht et al 2011,). Recently, a vigilance regulation model of affective disorders has been put forward (Hegerl & Hensch, 2014) in which behavioral syndromes are interpreted as an autoregulatory reaction to a dysregulation of vigilance. For example, hyperactivity and sensation seeking observed in mania and to a lesser degree in ADHD are interpreted as an autoregulatory attempt of the organism to stabilize vigilance by increasing external stimulation (in analogy to the behavior exhibited by overtired children). On the other hand, the withdrawal and sensation avoidance seen in major depression (MD) is interpreted as a reaction to a hyperstable vigilance regulation. Indeed, compared to healthy controls, a more stable vigilance regulation was reported in unmedicated patients with MD (Hegerl et al., 2012, Olbrich et al., 2012b; Ulrich & Fuerstenberg, 1999) and there are several hints for an unstable vigilance regulation in ADHD and mania (Hegerl et al., 2009/2010; Sander et al., 2010; Schönknecht et al., 2010; Small et al., 1999.

To facilitate research on vigilance regulation, a computer algorithm has been developed by the neurobiological research group led by Prof. Ulrich Hegerl at the Department of Psychiatry (University Hospital Leipzig, Germany). VIGALL 2.0 (Vigilance Algorithm Leipzig, VIGALL[1]) is an EEG- and EOG-based algorithm which enables to objectively assess the level of vigilance within EEG recordings, by automatically attributing the respective vigilance stage to EEG-segments (Hegerl et al., 2008; Olbrich et al., 2009, 2012a, Sander et al., submitted). This allows investigations on how vigilance is regulated during the recording period. The algorithm used in VIGALL 2.0 takes into account both the frequency patterns and the cortical distribution of EEG activity using EEG source localization approaches (Low Resolution Electromagnetic Tomography, LORETA; Pascual-Marqui et al., 1994/2002). Since EEG activity is characterized by high intra-individual stability and large inter-individual variability, VIGALL 2.0 has adaptive features concerning individual alpha peaks and amplitude levels.

VIGALL 2.0 improves upon an earlier version of the algorithm which has been validated performing simultaneous EEG-fMRI- (Olbrich et al. 2009) as well as simultaneous EEG/FDG-PET-studies (Guenther et al 2011) and by relating the vigilance stages to different behavioral and autonomic parameters (Minkwitz et al 2011, Olbrich et al. 2011).

VIGALL 2.0 is not applicable for certain EEGs, e.g. those showing alpha variant rhythms, major modifications due to drugs (e.g. anticholinergic drugs) or diseases (e.g. severe Alzheimer’s disease) or to EEGs from children under the age of 10 (or older in case of delayed maturation.

Vigilance levels and their regulation strongly influence the outcome to all kinds of tasks (e.g. reaction time, sustained attention, all kinds of cognitive tasks) and neurobiological recordings (ERP, functional brain imaging), and are part of the pathophysiology of most neuropsychiatric disorders. Concrete examples for application of VIGALL are:

• To monitor changes in wakefulness during resting-state MRI recordings. Such monitoring should not be limited to switches between sleep and wakefulness (Tagliazucchi & Laufs, 2014) but should also take into account that the wake state itself is also not homogeneous. According to a simultaneous EEG-fMRI study applying VIGALL (Olbrich et al., 2009), BOLD signal increases in several cortical and decreases in subcortical areas during vigilance declines.
• To study the relationship between pre-stimulus vigilance stage and reaction time or event related potentials. VIGALL allows to specify a certain time window around defined EEG-markers (e.g. markers indicating stimuli or reaction by subject), for which VIGALL will then classify the respective vigilance stage.
• Vigilance regulation provides a neurophysiological parameter for the study of wakefulness (with more detailed information compared to the Multiple Sleep Latency Test) in the context of psychiatric disorders or clinical trials. The routine clinical EEG recordings (~3min of closed eyes) do not allow studying this important neurophysiological marker. Prolonged (15min and more) resting-EEGs are needed which can be easily and rapidly evaluated using VIGALL (Hegerl et al., 2008/2012; Olbrich et al., 2011/2012a/b/2013).
• To control for differences in vigilance regulation in diagnostic tests that are dependent on brain activity and arousal. Up to now, such differences are rarely studied and in most cases not taken into account when interpreting group differences. However, doing so is likely to reduce variance and to improve diagnostic validity of functional brain imaging. For example, more than 50% of inter-individual variance in FDG-PET could be explained by inter-individual differences in vigilance stages during the trapping period (Guenther et al 2011).

Sincerely yours,

Ulrich Hegerl & Christian Sander and the Neurobiological Research Group

Department of Psychiatry and Psychotherapy
University of Leipzig
Semmelweisstr. 10, 04103 Leipzig, Germany
Ulrich.Hegerl@medizin.uni-leipzig.de
www.uni-leipzig.de/~vigall/