"Doc why couldn't you rush down early? He wouldn't have been as much critical as he is now!"
"Oh! come on. The patient is senseless. And I can't read mind!"
These conversations would never ever be taken place anywhere in the earth with KnowPain solution. Don't trust me? watch this video Video of KnowPain Prototype demonstration
Platform : Tablet
Programming Language: C#
Additional Hardware: Arduino based EEG board (2 Models will be sent for Judging)
Project Status: COMPLETED (Polishing Needed in UI and Communication Module)
What the App Does?
KnowPain is basically a complete package that includes a small hardware powered by Arduino to read, filter and measure mind signal called EEG signal. This is a hardware board which is designed by me that can track the regular EEG signal between two hemisphere of brain and that can track delta activity in EEG signal. If a person suffers from pain ( even if he is in sleep or senseless), the hardware tracks that. In is interfaced to the Tablet using USB port. A UI runs in the tablet that continually monitors the EEG signal coming from subject's brain and performs filtering. The filtering is high passed filter which is done by filtering out low frequency signals, that are nothing but regual EEG wave and passing only very high frequency waves ( the delta activity). This high frequency wave appears in EEG when a subject is under tremendous trauma or pain.
( Generally any feeling is an electrical signal that is propagated through nerves. In case of pain, there is a sudden surge in pulses. wherever the pain occurs, is transmitted urgently to brain, increasing the amplitude of the signal. One simple example is when you are hit by something, you see a blackout. It is called Concussion which is caused by temporary change in the way brain functions.)
The signal is processed by KnowPain app running in the tablet and as soon as a pain is detected, shoots an alarm, ( either as hardware buzzer or speech synthesis voice) and also sends an immediate SMS to specified Number, or can generate and send a bluetooth message or can send a push mail). It informs the doctor or medical caretaker so that immediately patient can be attended.
Every year around the globe, more patients dies due to medical negligence more than ill treatment or lack of it. Several cases of negligence are reported mainly in ICU and children care units. Cases are more eminent in the night where in many nursing homes attenders and nurses gets relaxed or sometimes decide to take a quick nap. Outcome needs no elaboration. So KnowPain is a solution that keeps the attender alerted all the time in emergency.
India has a pathetic doc/patient ratio(about 1 doctor par 1953 patients) . The ratio, needless to say a makeup. In reality, in villages, there is a single doctor for almost six villages ( Who cares about poor patients here?). So managing health care is extremely difficult. Maternity care, child care, care for stroke or other patients takes a spin here. Hence KnowPain could be next wanna be innovation in healthcare.
The hardware we propose here can be constructed at a very low rate ( <$40). Therefore it can be provided to almost all the villages and nursing homes. The UI can push data or alert through different possibilities including GSM (Make no mistake. Mobile and Television is available almost at every village.) Electricity is a big problem in India where barring few states, power supply at best is on the mercy of God. Admitted, most of the nursing homes have generator facility, but many don't have. And talking about villages, you certainly don't have that. To my understanding, the tablet should have better battery backup than the laptops. The processing does not really require very high CPU utilization due to low graphics task. Filtering is implemented at 1024 point FFT. No IFFT. Hence at any instance processing power is bare minimum. Arduino boards can be externally powered using 9v battery. Therefore all in all power drawn by the Hardware and Processing unit would be minimum. If Tablet comes with 3G facility, SMS push could be activated from the device, otherwise any bluetooth enabled mobile can be used as GSM modem for sending message. A typical EEG acquisition device is quite costly and does not come with any push alert system.
How medically authentic the device will be?
Remember we are not proposing a diagnosis device here. We are proposing a device that can identify pain. As you can see in the video, when the subject pinches himself the EEG detects it. With my budget, I can not get a medical authority certification for the device but it would not need one either. The electro isolation ensures that no reverse voltage goes to our body. We have designed high impedence load, which will efficiently draw the current from head EEG electrodes to Arduino board. Body exposure will be less than .05mA which would follow the medical norms. Hence the device would be safe to handle. Yes but for commercially using it, we need more hardware polishing and got to have some medical licensing, which can be obtained if ever the app gets acceptability. So it will be safe and will detect pain signal with an accuracy of 77% ( as par our current tests).
What is the development approach?
1. Arduino based hardware board that would acquire EEG signal, as demonstrated in Video.
2. C#.Net program would be acquiring the data from Arduino using serialPorts.
3. FFT unit will perform FFT over data.( You can read more about FFT and Filtering here)
4. Hamming window will be applied after using a bandpass filter to reduce stop band noise.
5. Once pain is detected, speech synthesis will keep announcing the problem. SMS unit will push the SMS to specified doctor's or attenders mobile.
Will it be completed in time? What is the proof of concept?
Biomedical engineering is one of my research areas and I have hacked almost every device and build many including Blood Glucometer , Pressure and Temperature Monitoring Device , ECG Processing .
I have written and shared several open source codes in this direction, especially in ECG and EEG signal processing including my codeproject article for ECG feature extraction .
If the above videos and work of mine fails to convince you that this path breaking product can really be delivered in no time by me than watch this video to test the concept by yourself. I hope this nail the issue that if there is a single App that is needed in healthcare now for developing countries, it is KnowPain