Towards a neurobiology of dysfunctional arousal in depression: the relationship between beta EEG power and regional cerebral glucose metabolism during NREM sleep

Abstract

This study sought to clarify the neurobiological basis of variations in one aspect of central nervous system ‘arousal’ in depression by characterizing the functional neuroanatomic correlates of beta electroencephalographic (EEG) power density during non-rapid eye movement (NREM) sleep. First, nine healthy (n=9) subjects underwent concurrent EEG sleep studies and [¹⁸F]2-fluoro-2-deoxy-d-glucose ([¹⁸F]FDG) positron emission tomography (PET) scans during their first NREM period of sleep in order to generate hypotheses about specific brain structures that show a relationship between increased beta power and increased relative glucose metabolism. Second, brain structures identified in the healthy subjects were then used as a priori regions of interest in similar analyses from identical studies in 12 depressed subjects. Statistical parametric mapping was used to identify the relationship between beta power and relative regional cerebral glucose metabolism (rCMRglu) during NREM sleep. Regions that demonstrated significant correlations between beta power and relative cerebral glucose metabolism in both the healthy and depressed subjects included the ventromedial prefrontal cortex and the right lateral inferior occipital cortex. During a baseline night of sleep, depressed patients demonstrated a trend toward greater beta power in relation to a separate age- and gender-matched healthy control group. In both healthy and depressed subjects, beta power negatively correlated with subjective sleep quality. Finally, in the depressed group, there was a trend for beta power to correlate with an indirect measure of absolute whole brain metabolism during NREM sleep. This study demonstrates a similar relationship between electrophysiological arousal and glucose metabolism in the ventromedial prefrontal cortex in depressed and healthy subjects. Given the increased electrophysiological arousal in some depressed patients and the known anatomical relations between the ventromedial prefrontal cortex and brain activating structures, this study raises the possibility that the ventromedial prefrontal cortex plays a significant role in mediating one aspect of dysfunctional arousal found in more severely aroused depressed patients.

Introduction

Depressed patients are often characterized as having abnormal levels of ‘arousal’, be it hypo- or hyperarousal. The clinical correlates of arousal may include hypersomnia or insomnia, increased or decreased appetite, and psychomotor retardation or agitation. In stress research paradigms, this increased central nervous system arousal is necessary in a healthy reaction to a stressor, yet in the case of depression, there is a hypothesized inability to terminate this arousal response following the removal of the stressful situation (Gold et al., 1988, Chrousos and Gold, 1992). Neurobiological studies have provided some support for the clinical construct of alterations in arousal using more operationalized definitions of arousal depending on the brain system or function under observation. Disturbances in both central neuroendocrine function and in sleep have been used to support the notion that some depressed patients have abnormal central ‘arousal’ in depression (Kirkegaard et al., 1975, Matussek et al., 1980, Gold et al., 1988, Benca et al., 1992, Chrousos and Gold, 1992, VanCauter, 1994, Nofzinger et al., 1999a). EEG sleep measures of abnormal physiological arousal are among the more reliable and well-studied of biological alterations in depression, and these measures have been linked with the clinical course of a depressive episode. Disturbances of EEG measures of sleep in depressed patients reflecting a more aroused pattern have predicted poorer response to psychotherapy and an increased likelihood of recurrence (Giles et al., 1987, Rush et al., 1989, Jarrett et al., 1990a, Jarrett et al., 1990b, Thase et al., 1997a, Thase et al., 1997b). These studies suggest that the identification of the functional neuroanatomical correlates of these shifts in cortical function to a more aroused state may help to define the brain systems that relate to clinical response in depression.

EEG power spectral density measures reflect the degree of global cortical electrophysiological arousal. Higher frequency, lower amplitude EEG waves such as those seen in the beta frequency range reflect the desynchronized, activated cortex most pronounced in waking and rapid eye movement (REM) sleep. Higher frequency EEG activity is associated with behavioral arousal and attentional processes (Tzischinsky and Lavie, 1994, Lamarche and Ogilvie, 1997, Wyatt et al., 1997, Merica et al., 1998, Nofzinger et al., 1999b). High frequency beta power spectral density is inversely related to power in the lower frequency, higher amplitude delta band that characterizes cortical deactivation and that is associated with deeper NREM sleep (Merica and Fortune, 1997, Portas et al., 1997). Interestingly, a temporal coupling between EEG power in the beta frequency and cortisol secretory pulses have been shown (Chapotot et al., 1998) suggesting a mechanistic link between increased function of the hypothalamic-pituitary-adrenal (HPA) axis and higher frequency brain activation. These observations suggest that the brain mechanisms that are associated with variations in beta EEG power may play a role in mediating an aspect of the clinical variations in arousal seen in depression.

Within this framework, we assessed both EEG power spectral density and regional cerebral glucose metabolism using [¹⁸F]2-fluoro-2-deoxy-d-glucose ([¹⁸F]FDG) PET scans during sleep in healthy and depressed subjects. We focused on the first NREM sleep cycle since this period is, in general, the time when brain arousal is at a minimum. Persistence of higher frequency EEG activity at this time may suggest the presence of a more aroused brain. In the current study, we focused on the higher frequency beta band given its inverse relationship with delta power spectral density (Merica and Fortune, 1997, Portas et al., 1997) and its preclinical relationship with HPA-axis function (Chapotot et al., 1998). Statistical parametric brain maps (SPM) were generated defining regions of the brain that covaried positively with the EEG power spectral density in the beta frequency range as one indicator of electrophysiological arousal. Given limited information in this area, studies in healthy subjects were conducted without a priori hypotheses. Regions identified in healthy subjects were then used in hypothesis-driven studies in the depressed group. In order to more accurately identify brain structures localized by PET scanning, we performed brain magnetic resonance scans for co-registration with the PET images. Baseline EEG sleep studies and subjective rating scales of sleep quality were also assessed in order to characterize the relationships between objective and subjective measures of arousal during sleep.