The global digital migraine treatment devices market is projected to reach $2.3 billion by 2033. Theranica is one of the largest players in the neuromodulation segment. It develops Nerivio, an upper-arm wearable for acute and preventive migraine treatment. Last week, the company published an analysis of 770,473 daily reports from 53,065 Nerivio users.
Published in Neurology Open Access, the study tested machine-learning models incorporating headache history, symptoms, and other data collected through the Nerivio app. The best-performing model could predict whether a migraine would occur the following day with 91.2% accuracy. The findings help Theranica shift migraine treatment from reactive to preventative.
Theranica’s study drew on data collected through the Nerivio app between January 2020 and July 2025. Of 58,863 users who had reported symptoms at least once, 53,065 met the study criteria, producing 770,473 evaluable target days. The dataset combined electronic migraine diaries and pre-treatment questionnaires with demographic information, local weather data, and derived measures such as recent migraine frequency and headache severity.
The researchers tested seven machine-learning approaches to predict whether a user would report a migraine on the following calendar day. An optimized XGBoost model performed best overall, reaching 91.2% precision, 81% accuracy, 80% sensitivity, 83% specificity and an AUC of 0.893. In practice, that means that 91.2% of days flagged as migraine days were followed by a reported migraine, while the proportion of all days classified correctly was 81%.
Much of that predictive power came from longer-term headache patterns. Average headache severity over the previous 30 days was the model’s strongest individual feature, contributing 37.6% of feature importance. Migraine frequency over the previous seven days contributed another 9.6%. Features derived from the full 30-day period together accounted for 56.3%, while information from the current day contributed just 4.4%.
The results suggest that a patient’s recent migraine trajectory may provide useful warning before an individual attack. Theranica has already incorporated forecasting into the Nerivio app through its Your Day Ahead feature, which provides an estimate of next-day migraine likelihood.
Nerivio uses an approach Theranica calls Remote Electrical Neuromodulation, or REN. The wearable sits on the upper arm and delivers electrical stimulation to peripheral nerves, aiming to activate the body's descending pain-control pathways. In that sense, it differs from the majority of migraine devices that stimulate around the head or neck.
The device first received FDA De Novo authorization in 2019 for acute migraine treatment in adults. Its indication has since expanded to both acute and preventive treatment of migraine, with or without aura, in patients aged eight and older. Preventive use involves a 45-minute session every other day, while acute treatment activates at the onset of symptoms.
Clinical studies have supported both uses. In a randomized trial of acute treatment, 66.7% of patients receiving active REN achieved pain relief after two hours, compared with 38.8% receiving sham treatment. A later preventive trial found that patients using Nerivio every other day reduced monthly migraine days by 4.0 days, compared with 1.3 days in the sham group.
Theranica sits within a broader migraine neuromodulation market spanning trigeminal nerve stimulation, vagus nerve stimulation, and implanted approaches for harder-to-treat patients. Nerivio had been used in more than 2 million treatments by May 2026, while US insurance coverage reached 158 million lives in July.
.svg)
