Ayurveda Wearable: Why go beyond HRV data?

My journey of development of the #personalized, #affordable and #accessible wearable, based on Ayurvedic principles, started with this first conclusive statement from my 16+ years of research: The Heart Rate Variability (HRV) data is not sufficient; whether it comes from optical sensors of smart watches or fitness gadgets, or from mobile camera sensors of mobile phones, or from plethysmograph, or from sensors of oximeter, etc. In this article, I am going to elaborate on this first question from my previous article "Ayurveda Wearable: Not just fitness, but your journey to Wellness".

The image (a) shows two different biomedical signals with respect to time. One is a typical pulse signal (black) from our patented sensor with 0-4 inch H2O pressure range. Second one is a typical pulse wave signal (red) from optical/mobile camera sensor, we all have observed similar waves in oximeters also during the pandemic period. Observe the marked blue regions.

????? ???? ?? ???????? ????? ????? ...

It is clearly visible that the red data is not sufficient to capture the "pressure" information on the wrist which has been defined and studied for the last 5000 years in Ayurveda, where a Vaidya with concentration and experience feels the pulse and look for information based on -- depth, intensity, amplitude, frequency, rhythm, length, type, temperature, quantity, texture and width. This information is then converted/ thought as various gati such as is the pulse feeling similar to movement of a snake (Sarpa gati) or jumping similar to frog (Manduka gati) or swinging smoothly like a swan (Hansa gati)... Its a long list of different gatis and each gati signifies certain amount of presence of Vata, Pitta and Kapha dosha and also further related to other gunas, dhatus and so on ...

VPK information based on Frequency analysis of the HRV data:

For getting the Vata, Pitta and Kapha information from HRV data, frequency domain properties are used. There are 100s of technical papers available online, the simplest one I have come across is by Dr. Ram and Dr. Oleg (Link: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6033724/). It shows correlations of low, medium and high frequency ranges in the HRV with VPK. Reading this and many other papers, it is evident / important to collect the HRV data for longer duration (few minutes for sure) according to me to actually provide information about the low and high freq components. Also, it totally misses out the information of other features from the above list such as depth (maybe), texture (surely), type (maybe) and so on...

[Best paper for the complete understanding of HRV: Heart rate variability, European Heart Journal (1996) 17, 354–381. And for the deeper understanding the nonlinear Poincaré properties, go through Filtering Poincaré plots, Computational methods in science and technology 11(1), 39-48 (2005)]

Pressure Information

When I started my PhD at IIT Bombay working on this, Vaidyaraj Ashok Bhat (Pune) was the domain expert. He used to teach his students about the "feeling information" of the pulse through various objects. One of such objects is shown in the figure (b), different portions of the knife have different feelings when touched and the pulse on the wrist when "felt very carefully", provides such different information patterns. And the information available in 5 seconds / 15 seconds / 1 minute of data is so vast that -- how many times Kathinata was available, how many times Sukshmata was available, how many times Sukshmata in Kathinata was available -- this information in the pulse is quite complex and convoluted. These Sukshma, Tikshna information sets are very important for understanding the gatis and inner health information of the subject. And is available only when the data is acquired using a very sensitive pressure sensor - completeness of data.

I remember, when I was working on various sensors as part of my research; I used to collect data at Vaidyaraj Bhat's clinic or at Prof. Sharat Chandran's lab in IITB and we used to discuss what pattern is achieved and what is missing. One day, I had a Eureka moment. So, I took a Nadi at IITB and showed him the print out of the Nadi (nothing else) at Pune. Just looking at the print, he said - this is a female subject, roughly 25-30 age group, she is having a chest pain, that chest pain is because of gases, ask her to do - - - -. I was happily amazed and frightened at the same time, because I had taken the Nadi of Meghana (now my wife) because she was having chest pain at that time. I thought the development of sensor work is finished AND now the work I have to do is to convert what "he saw in the print" as patterns, extract them and decipher in today's language. At that time, he had explained a few pointers on the printout in the direction of energy, rhythm, intra-variability (not inter) and texture in his own language of 5 parameters.

Self-similar Chaotic Patterns in the Pulse

Under the guidance of Dr. B. D. Kulkarni and Dr. V. K. Jayaraman at NCL Pune, I extracted the topological invariants and chaotic parameters of the pulse to extract its self-similar nature (one of the results in figure c). The 3D attractor is achieved from a one dimensional pulse signal and is observed for its stretching and squeezing information -- to extract the conservativeness or fractal nature or predictability of the overall system. This research and outcomes help us in the early detection and predictions. We have observed that the repetitive or non-repetitive behaviour of various features actually tells a lot of information of the Nadi gunas. This important information is completely missed in the pulse wave signals from other sensors.

In the last 200+ years, heart signal and HRV-based signal has provided ways to get information about the cardiovascular issues, let us now go beyond -- Personalized / Patient-centric, Preventive, Predictive, Pervasive and Participatory !! In my PhD thesis "Acquisition and Quantitative Analysis of the Arterial Pulse", we talked about repeatability, reproducibility and most importantly completeness of the data. A few companies have millions and billions of the data from the smartwatch or fitness gadget, but still the personalization factor is yet to be solved...

In Die Hard series, John McClane was a wrong person at the wrong place at the wrong time.

Atreya's Ayurveda-based wearable will have the right "sensor", at the right "place", at the "right" time; but may not be only on the right hand :)