Introduction

Seventy percent of head and neck cancer - the sixth most common malignancy - patients have oropharyngeal cancer (OPC), caused most commonly from human papillomavirus (HPV) infection, the most common HPV-related malignancy in the US. Radiation, the mainstay OPC treatment, can induce neuropathy along the cranial nerves (glossopharyngeal, and hypoglossal) with a 10-48% incidence, resulting in swallowing deficits. RICN results in dysphonia, dysphagia, loss of taste, and tongue fibrosis. Swallowing deficits can impair quality of life and can even result in aspiration. Although clinical trials show that steroids (NCT04151082) and gabapentin (NCT03747562) temporarily reduce pain, their adverse effects compromise tolerance and adherence. Thus, we need new treatments.

Here we strengthen networks that regulate swallowing through our individualized fMRI neuromodulation, termed iNM (U.S. Patent No.16/954,256). fMRI measures the magnitude and spatial extent of the ratio of oxygenated (O₂) to deoxygenated hemoglobin [Hb]. (1) iNM is a non-invasive, precision-medicine intervention that can strengthen primary motor and sensory areas that regulate swallowing in the brain with 1mm precision without interfering with other treatments or medications, (2) iNM targets each patient’s unique anatomical and functional circuitry that regulate swallowing, and (3) iNM is guided by reinforcing or inhibiting the HbO2 intensity and extent of each patient’s unique brain network, as opposed to the self-regulation of the HbO2 intensity. We decode cortical selectivity for swallow to elucidate the mechanisms of iNM and to aid in the development of more sensitive and specific personalized neurorehabilitation in OPC patients with RICN suffering from dysphagia.

Methods

Healthy subjects (n=30) participated in a two-day iNM study. On Day one, we decoded the brain spatial patterns and amplitude generated by swallow, interleaved with periods of tongue-rest(baseline) and swallow. Our innovation is based on targeting individualized networks by extracting HbO2 magnitude and spatial extents. Each participant’s TMSC cortical selectivity was extracted and targeted for iNM. On Day two, participants underwent iNM and control-NOiNM conditions. Support vector machine (SVM) classified cortical swallowing patterns versus tongue-at-rest generated via iNM and control conditions. We quantified the HbO2 magnitude for each network’s areas separately, by computing dynamic causal modeling to elucidate the association between brain states and the physiological responses of TMSC. SVM modeling was trained in 29 participants and tested on 1 participant an iterative fashion using a sliding window.

Results

iNM resulted in higher classification accuracy compared to the control condition. Enhanced activity was noted for motor cerebellum and basal ganglia areas. Neural proprioceptive and pain matrix substrates, such as the insula and claustrum, were activated under control but not iNM. 

Conclusion

iNM can strengthen the physiologic response of swallow and increase the signal-to-noise ratio by reducing the network spatial extent. iNM reinforces greater control during the oral preparatory phase by inhibiting pain areas, such as the insula and claustrum, and enhancing motor control areas. Understanding the mechanisms of swallow under iNM will help us delineate optimal neuro-rehabilitative targets for OPC RICN patients suffering from dysphagia.