What is SeizAlert?

SeizAlert alerts the wearer and medical personnel of an impending epileptic seizure, using is a low-cost, compact, wearable device. SeizAlert forewarning allows preventive action, reduction in morbidity and mortality, and improvement in patients' quality of life. SeizAlert offers a new treatment paradigm of constant monitoring, rather than continuous medication. Reliable, long-lead-time forewarning allows the patient to stop hazardous activity, lie down in a quiet place, undergo the seizure, and then return to normal activity. Other timely preventive steps may include taking medication, requesting emergency responders, and/or contacting the physician.

How does SeizAlert work?

The SeizAlert technology presently uses brain-wave data from four scalp electrodes at the front of the head, removes eye-blink (and other muscular) artifacts that would otherwise confound the analysis, converts the artifact-filtered data into a statistical distribution function (DF), and compares the DF for the baseline period with subsequent DFs via measures of dissimilarity. Several successive occurrences of the dissimilarity measures above a threshold provide up to 4.5 hours forewarning of an impending event with a total true rate of 56/60 (93%). The approach requires an off-line, analyst-intensive optimization of the processing parameters, after which the analysis runs much faster than real time on a 1.2 GHz palmtop PC (analysis of 261 wall-clock hours of brainwave data in one hour). SeizAlert is covered by eight US patents. The first paper under Biomedical Analysis on the Publications pages provides technical details: http://computing.ornl.gov/cse_home/staff/hively/publications.shtml. PDF copies of the SeizAlert patents are also available on the Publications page under US Patents.

What is the status of the SeizAlert technology?

R&D Magazine announced that SeizAlert is a 2005 winner of its international-class RD100 award. ORNL is seeking industrial partner(s) to license the technology.

Can the SeizAlert technology forewarn of other biomedical and machine events?

Yes, ORNL has demonstrated that the SeizAlert technology also provides forewarning of other biomedical and machine events. Additional biomedical applications to date include: forewarning of heart attacks (ventricular fibrillations) and fainting (syncope) from human heart-wave data; detection of septic shock due to inhaled endotoxin from rat heart waves; and detection of breathing difficulty from pig chest sounds. Other potential biomedical applications include remote monitoring of soldiers, emergency responders, etc. Machine applications to date include forewarning of failures in electric motors (air-gap offset, broken rotor, turn-to-turn short, and imbalance); motor-driven components (gears, bearings, pumps [imbalance and misalignment], rotating blades, drill bit wear) using a variety of process-indicative data (torque, vibration, electrical power, voltage, current); and forewarning of structural failure by cracking from time-serial stress and strain data. The first (review) paper on the Publications page provides technical details about these biomedical and machine applications: http://computing.ornl.gov/cse_home/staff/hively/publications.shtml.

Whom can I contact for more information about SeizAlert?

If you have technical questions, please contact Dr. Lee Hively (hivelylm@ornl.gov; 865-574-7188).
If you have questions about licensing of the SeizAlert technology, please contract Dr. Greg Flickinger (flickingergc@ornl.gov; 865-241-9485).

• Publications
• PhD Thesis