Fatigue is a symptom of many health ailments. It is also a precursor of illness and an effective monitor of treatment as the body functions ineffectively with weakness. Yet, checking fatigue is underutilized in healthcare partly because questionnaires – the prevailing method – are subjective and difficult to implement.

By separating fatigue into physical, mental and sleep deprived sources of fatigue, gold standards were used to identify groups of pulse wave features best correlated to these more narrowly defined fatigue types. The Company is unique in using its patent-protected machine learning analysis to correlate fatigue and stress with pulse wave analysis with correlations of around 90%. (With additional data the algorithms are further “learned” and more features can be added to further increase accuracy). Using pulse wave forms and their variations, this methodology is not to be confused with heart rate frequency-based analysis – using one single feature – prevalent in “wearables”.

As noted in the table below, the Company’s pulse wave based algorithms can also discriminate between types of fatigue. With each additional feature added, the accuracy is increased:


Physical vs others

Mental vs others

Sleep vs others

One feature model
Two features model
Three features model

0.833 ± 0.095
0.896 ± 0.053
0.911 ± 0.054

0.746 ± 0.102
0.795 ± 0.067
0.842 ± 0.088

0.564 ± 0.086
0.766 ± 0.088
0.776 ± 0.081

Below are graphs illustrating the levels of fatigue quantified per fatigue: physical, mental and lack of sleep:


Click here for details on the study parameters for this fatigue study