Seeking Useful Biomarkers for the Quality and Effectiveness of Sleep

It has always been a source of disappointment that despite the richness of signals in the average polysomnogram, across a number of biological subsystems such as electrocortical, respiratory, hemodynamic, autonomic, and motor, we (clinicians and researchers in sleep medicine) have little software means to assess interactions and integration of signals. For example, a patient with sleep apnea, metabolic syndrome, endothelial dysfunction, and depression becomes an “AHI.” Several contributors to this reductionism can be recognized: (1) the prevalent scoring rules that offer actual punishment for thoughtful deviations (e.g., scoring discrete desaturations starting with a low baseline); (2) the single-channel or physiological stream scoring strategy (e.g., arousals are EEG phenomena, and 3 seconds only!); and (3) the disincentives to vendors and physicians to engage in new methods of data analysis. With the likely coming drop in reimbursement, it seems a bad idea to want to do more with what we have.

There could be seen an inherent conflict between the moves to small devices measuring essential information in the home environment versus extending the laboratory assessment to generate ever deeper insights into the process of normal and pathological sleep. The paper by Chervin and colleagues¹ in this issue of SLEEP is an example of an approach that has the potential to bridge that gap, and even allow information transfer across that bridge. This is one of a series of papers describing coupling of electrocortical activity to respiratory cycles, the respiratory cycle-related EEG changes (RCREC). The respiratory cycle is broken down into early and late inspiration and expiration, and the variance of the typical sleep EEG spectral energy bands across these segments are computed and averaged, the target being non-apneic breaths.² Thus, the technique seeks to assess the impact on electrocortical activity of partially obstructed airflow, and the associated physiological consequences (i.e., increased respiratory effort, afferent stimulation from the upper airway, intrathoracic pressure and volume changes, blood gas changes). Previous papers have shown correlations of specific signal bands with clinically important outcomes such as objective sleepiness,³ but this is the First to show the effect of positive pressure treatment. Briefly, Chervin et al.¹ found reductions in this evoked activity with application of therapy, with the contrasts made in the second half of the night between baseline and treatment trials, on a single night.

The results and technique itself begs a number of interesting and important (and mostly unanswered) questions. What generates RCREC? Can it be simplified such as only say one inspiratory phase is adequate? We do not know fine-mapped regional differences as the central derivation was used; it is entirely plausible that generating the RCREC from different and far smaller cortical volumes using bipolar electrodes, such as with a dense array EEG, may provide a dynamic and informative “RCREC map.” Nevertheless, RCREC may be generated from corollary rostral-oriented discharges from the respiratory neuronal groups, cranial transmission of peripheral respiratory sensation, and outflow from the solitary nucleus that can integrate chemoreceptor, baroreceptor, and other visceral afferents. In fact, low frequency stimulation of the solitary tract/nucleus area has synchronizing effects of NREM sleep.⁴ It seems that there is normal RCREC, and abnormal RECEC.

What could modify RCREC? While the paper by Chervin et al.¹ has a focus on the information from below, it is unlikely that the cortex will be a passive partner. The sleep EEG is enormously complex, with strong individual/genetic differences in frequencies and amplitudes,^5,6 age-related changes, NREM sleep periods dominated by phasic EEG activity or not,⁷ homeostatic and circadian influences, and discrete phenomena such as the K-complexes that impart large amounts of instantaneous delta power. Would a “depressed” brain, or say, one exposed chronically to alcohol, have different RCREC values and norms? Moreover, there is an ever increasing list of medications that can modify every sleep-related EEG frequency band (e.g., it is likely that RCREC would be affected by hypnotics).

What could be the natural course of RCREC over time? Clearly adaptive and maladaptive changes occur across biology in response to chronic stressors. The brain in the average patient with moderate or greater severities of sleep apnea has been under an intense and varied biological stressor for all the time the apnea was untreated. To think that just opening up the airway and everything will be all right is naïve—one reflection of that being residual sleepiness despite even perfect sleep-breathing management.⁸ Could RCREC provide a new window into understanding remodeling of cortical dynamics over time?

What could be the minimum unit of RCREC that can be biological relevant and clinically useful? What is the minimum number of averages to obtain robust responses, allowing a moving average across the night? The minimum numbers of respiratory cycles that require averaging to generate a reliable RCREC may vary across the time of night, sleep stage, with age, pathological states, and of course drugs. I would prefer a term other than microarousals to describe RCREC changes, as some of these changes are surely normal and some RCREC is always present.

The RCREC is one example of integrated cross-physiology metrics that may be generated from sleep. The numerous components of sleep state are part of a complex, dynamic, and interacting system. Some examples include the invariant surge in blood pressure in response to an arousal from sleep; motor activation triggered by respiratory arousals; and using the coupling between autonomic and respiratory activity to map sleep state.⁹ So we now know that respiration also couples with electrocortical activity.

The RCREC may be considered one of several candidate metrics to map the quality and effectiveness of sleep, including sleep stages, actigraphy, oximetry, pulse transit time, and blood pressure. Effectiveness is a term to designate sleep state and processes that allow the normal functions of sleep. Effectiveness and ineffectiveness may coexist in separate physiological traces (e.g., in a patient with hyperarousal-related sleep fragmentation, oxygenation would be expected to be normal). Each marker will have its unique place based on the clinical question. At the deepest level, the quality and effectiveness of sleep must engage core biological processes such as gene expression and epigenetic regulation. For example, hypoxia is one of the strongest epigenetic modifiers.¹⁰

I am less optimistic than Chervin et al.¹ when it comes to “what next.” I think the current zeitgeist in clinical sleep medicine may not encourage “incorporation of automatic RCREC analysis into real-time monitoring might improve effectiveness of diagnostic polysomnography, laboratory-based PAP titrations, or automatically adjusting PAP machines.”¹ I am inclined to believe that the future of such elegant analysis techniques is in the home, using small disposable sensors and computation on personal devices such as smartphones, enabling the patient and physician with unparalleled views of dynamic physiology across a life time of therapy.

DISCLOSURE STATEMENT

Dr. Thomas is a co-inventor and patent holder of the ECG-spectrogram technique.

CITATION

Thomas RJ. Seeking useful biomarkers for the quality and effectiveness of sleep. SLEEP 2012;35(2):173-174.