A new EEG biomarker of neurobehavioural impairment and sleepiness in sleep apnea patients and controls during extended wakefulness

Highlights

• We analysed resting, awake EEG data using two qEEG methods, power spectral analysis (PSA) and detrended fluctuation analysis (DFA), and assessed their ability to yield EEG biomarkers of neurobehavioural impairment and sleepiness.

• PSA and DFA biomarkers correlated with impaired performance and increased sleepiness, and baseline measures of the DFA scaling exponent, but not power spectra, were associated with impaired simulated driving after extended wakefulness in obstructive sleep apnea (OSA) patients.

• The DFA scaling exponent has potential as a useful biomarker of performance failure which can be used as an alternative to conventional power spectra.

Abstract

Objective

To explore the use of detrended fluctuation analysis (DFA) scaling exponent of the awake electroencephalogram (EEG) as a new alternative biomarker of neurobehavioural impairment and sleepiness in obstructive sleep apnea (OSA).

Methods

Eight patients with moderate–severe OSA and nine non-OSA controls underwent a 40-h extended wakefulness challenge with resting awake EEG, neurobehavioural performance (driving simulator and psychomotor vigilance task) and subjective sleepiness recorded every 2-h. The DFA scaling exponent and power spectra of the EEG were calculated at each time point and their correlation with sleepiness and performance were quantified.

Results

DFA scaling exponent and power spectra biomarkers significantly correlated with simultaneously tested performance and self-rated sleepiness across the testing period in OSA patients and controls. Baseline (8am) DFA scaling exponent but not power spectra were markers of impaired simulated driving after 24-h extended wakefulness in OSA (r = 0.738, p = 0.037). OSA patients had a higher scaling exponent and delta power during wakefulness than controls.

Conclusions

The DFA scaling exponent of the awake EEG performed as well as conventional power spectra as a marker of impaired performance and sleepiness resulting from sleep loss.

Significance

DFA may potentially identify patients at risk of neurobehavioural impairment and assess treatment effectiveness.

Introduction

Inappropriate sleepiness and impaired neurobehavioural function arise from insufficient or poor quality sleep or from sleep disorders such as obstructive sleep apnea (OSA) (Aloia et al., 2004, Engleman and Douglas, 2004). In OSA, these symptoms result from sleep fragmentation and chronic intermittent hypoxia (Beebe and Gozal, 2002) and have a major impact on health, wellbeing and the economy (Sassani et al., 2004, Hillman et al., 2006). OSA patients have a 2–10 times increased risk of motor vehicle crashes and at least double the rate of workplace accidents (Lindberg et al., 2001, George, 2007). Clinicians often face difficulty in identifying which patients are at risk of accidents because of the disparity between daytime symptoms and conventional metrics of disease severity [e.g. apnea hypopnea index (AHI)] (Beebe, 2005, Al-Shawwa et al., 2008, Quan et al., 2011). Moreover, the heterogeneity of impairment in the patient population adds to this problem – one patient may be relatively asymptomatic whereas another may be greatly compromised even though both have the same degree of disease measured by sleep study (Beebe, 2005). Currently, there are no simple, robust biomarkers that reflect the negative impact of OSA on the brain, as measured by cognitive outcomes and driving.

Extended wakefulness paradigms are used as a way to challenge individuals and expose those who are vulnerable to sleep-loss induced performance deficits, separating them from those who are seemingly resilient (Mu et al., 2005). Previously, we used this experimental approach to assess differences in the response to sleep loss in OSA patients and controls. Contrary to expectation we found no difference in neurobehavioural performance (simulated driving and vigilance), between groups (Wong et al., 2008). Although performance deterioration was comparable, differences in brain activity are highly likely. Functional magnetic resonance imaging (fMRI) suggests that OSA patients may recruit different brain areas to counterbalance performance decrements and compensatory mechanisms may conceal expected differences in how OSA patients perform when compared to controls (Ayalon et al., 2006).

There is a clear need for a simple, practical tool that can be routinely administered during wakefulness to identify those individuals most susceptible to sleep loss modulated cognitive impairment. Of particular concern are those untreated OSA individuals in high-risk professions e.g. transport workers. One approach to address this issue is the use of quantitative analysis of the awake electroencephalogram (qEEG). One method of qEEG analysis is power spectral analysis (PSA). PSA metrics of the awake EEG have been shown to correlate with sleepiness and worsening neurobehavioural performance in healthy individuals during sleep deprivation experiments (Kecklund and Akerstedt, 1993, Cajochen et al., 1995, Lal and Craig, 2002, Campagne et al., 2004). PSA has also identified group differences in awake EEG during 24 h of sleep deprivation with a higher mean power reported in OSA patients than controls (Grenèche et al., 2008). PSA does not however, appear to correlate with subjective alertness in OSA patients undergoing sleep deprivation and data on neurobehavioural performance were not reported in this study (Grenèche et al., 2008). There are also acknowledged limitations with PSA. It makes assumptions about the linear and stationary nature of the EEG, which are not its true characteristics (Seely and Macklem, 2004). Interpretation of results can be cumbersome due to the multiple power spectral metrics calculated for the different and often inconsistently defined frequency bands i.e. alpha, delta, low delta, theta, beta, sigma, etc.

A simpler alternative method to quantify EEG is detrended fluctuation analysis (DFA) which provides a single metric, called the scaling exponent. In contrast to PSA, DFA has the methodological advantage that it is appropriate for nonlinear and non-stationary physiological data such as EEG (Peng et al., 1995). For a given time interval, DFA integrates the EEG signal, and more easily removes artefacts (noise and trends) (Peng et al., 1995, Robinson, 2003, Kantz and Schreiber, 2004). The scaling exponent increases during the transition from wake to sleep, continues to increase with deeper stages of NREM sleep, and correlates with traditional PSA metrics (Kim et al., 2009). In analysing other physiological signals such as heart rate variability, DFA provides additional prognostic information compared with PSA (Penzel et al., 2003). It is possible therefore that DFA could identify differences in brain activity between OSA patients and controls that may help explain the unexpected results of our earlier work of comparable vigilance and driving during 40 h of extended wakefulness (Wong et al., 2008).

Using EEG data collected during our previous study (Wong et al., 2008) we analysed resting, awake EEG data using two qEEG methods: conventional PSA and the simpler alternative, DFA. We assessed their ability to yield EEG biomarkers of neurobehavioural impairment and sleepiness exposed during a 40 h extended wakefulness challenge. We speculated that PSA and DFA biomarkers, power density and scaling exponents respectively, would demonstrate comparable relationships with neurobehavioural performance and sleepiness levels, supporting the use of DFA as an alternative to PSA.

We hypothesised that (1) the EEG biomarkers would correlate with impaired neurobehavioural performance assessed by driving simulator and psychomotor vigilance tasks, and increased sleepiness in both groups; (2) baseline measures of the EEG biomarkers would reflect ‘at risk’ individuals who are more susceptible to impaired simulated driving with sleep loss during the circadian night-time or early hours of the morning, a period of increased vulnerability to performance decrements (Cohen et al., 2010); (3) qEEG analysis would identify group differences in brain activity across the extended wakefulness challenge between OSA patients and controls.