Brain biomarkers for pain sensitivity

Identifying objective biomarkers that track individual pain severity have been dubbed “the holy grail” of pain neuroscience. While pain is among the most fundamental, ubiquitous, and adaptive experiences that can befall an organism, there is still a murky understanding of how pain is generated in the nervous system. Modern consensus on brain mechanisms underlying maladaptive chronic pain (pain that persists for greater than 3 months) is even less clear. Chronic pain affects up to one-fifth of US adults and its complexity is attributed to a confluence of physical, emotional, and cognitive factors that contribute to suffering and disability^1,2. The epidemic of chronic pain initially contributed to the rise of the opioid epidemic and continues to plague nearly all fields of clinical medicine. Identifying and validating biomarkers to predict individual risk for chronic pain, facilitate a precision medicine approach to pain medicine³. A new human study using an experimental prolonged pain model by Chowdhury et al.⁴, may bring us one step closer to the objective prediction of individual pain sensitivity.

In this technically rigorous and innovative study, investigators identified two cortical biomarkers that predicted future sensitivity to prolonged temporomandibular pain. This large-scale cohort validation extends previous observations about the role of peak alpha frequency^5,6 and corticomotor excitability in predicting pain sensitivity. The authors used an experimental model of prolonged pain by performing two injections of Nerve Growth Factor (NGF) into the right masseter muscle of 150 healthy human participants who also provided twice daily ecological momentary assessments of ongoing pain for four weeks. Injection of NGF produced a wide range of reported pain intensities across individuals, with some reporting mild pain (3 out of 10) and other reporting very intense (9 out of 10) pain that slowly waned over 30 days. Using rigorous growth mixture models, the authors then assigned each participant to a low or high pain sensitivity group based on their individual pain trajectory. Two key biomarkers could predict whether each participant would go on to experience mild versus severe levels of jaw pain over the ensuing month. Importantly the two biomarkers performed classification better when used together than either one did alone, achieving a combination discriminability of 88% area under the receiver operating characteristic curve.

The first biomarker, central peak alpha frequency (PAF), was computed from 64 channel scalp electroencephalogram recordings of as little as two minutes duration, even before NGF injection or onset of pain. The second biomarker measured changes in cortical motor excitability (CME) obtained by applying single pulses transcranial magnetic stimulation (TMS) to the left frontal scalp overlying motor cortex between days 0 and 5. Specifically, participants showed either an expansion (facilitation) or contraction (depression) in contiguous left scalp area that elicited a motor evoked potential in the right masseter muscle, indicating either increased or decreased excitability as their jaw pain was peaking. For both biomarkers, lower values were associated with a greater likelihood of high pain sensitivity. Another critical aspect of rigor in this study was the use of an entirely separate cohort of volunteers as an independent test set for biomarker validation. These biomarkers were reproducible and reliable, demonstrating excellent future pain sensitivity prediction even when calculated with variations to the analytical methods. Finally, the authors convincingly demonstrated the value of the combination biomarkers, which performed better prediction of pain than alternative models that incorporated potential covariates such as sex or pain catastrophizing scores.

The authors interpreted PAF as representative of ascending spinothalamic input while CME was taken to reflect descending modulatory circuits arising from the motor cortex. Currently, we lack clear evidence illuminating the role of PAF or CME in physiological brain mechanisms that underlie chronic pain development or in the transition from acute to chronic pain. Furthermore, it is possible that these biomarkers are not causal for specific chronic pain syndromes but rather reflect some epiphenomenon of chronic pain mechanisms such as central sensitization (a known mechanism of chronic pain⁷). Further work understanding the causality of this link may lead to a better awareness of whether direct pharmacology or neuromodulation interventions can reduce future pain sensitivity. If we could speed up an individual’s peak alpha frequency or augment corticomotor excitability, could this reduce the risk of developing chronic pain?

Despite the unknown provenance of the proposed biomarkers, the results of the present study have the potential to aid in the practical development of personalized pain management strategies. The clinical utility of these biomarkers is especially promising due to their demonstrated generalizability and non-invasive implementation. Specifically, the proposed biomarkers are suitable for informing intervention strategies based on a prediction of future pain sensitivity - an individual level trait that may relate to the risk of chronification of clinical pain symptoms. Future personalization through adaptive pain management will be enabled by advancing our understanding of biomarkers that predict pain state (i.e., intraindividual moment to moment fluctuations in subjective pain intensity⁸), similar to recent work in disorders such as treatment resistant depression⁹ and obsessive compulsive disorder¹⁰.

We are excited about the validated biomarker presented in this paper, which will likely have broad applicability across many medical fields. If successfully translated into clinical practice, biomarkers that predict a transition to chronic pain would have tremendous impact for the treatment of millions of individuals. Beyond the individual results here, we expect that this paper will inspire further research in biomarkers validated in clinical populations rather than pain models, biomarkers of ongoing pain state, approaches that lower barriers to adoption (i.e., not requiring TMS stimulation or precise timing of measurement relative to pain onset), identification of physiological mechanisms, and operational use in a variety of settings (including identification of high risk individuals, tailored pain management strategies, or screening of new interventions). While progress in quantifying subjective percepts is exciting, these advances must be accompanied by concurrent advances incorporating global neuroethics guidance¹¹ and addressing the specific ethical concerns related to pain treatment¹². Regarding specific concerns arising from the use of pain biomarkers¹³, we must take care to ensure that quantitative measures do not supplant lived experience reports, introduce distrust in the doctor-patient relationship, set unrealistic patient expectations, or exacerbate existing inequalities¹⁴ in pain treatment across this vulnerable population.