Aug 2018 Financial Express
Where there are connected devices, can malicious hacking and cyberattacks be far behind? As the concept of intelligent healthcare gains acceptability, with wireless implantable medical devices such as cardiac defibrillators and insulin pumps being utilised by patients for a fresh lease of life, the danger of these devices being manipulated is ever-present. Implantable devices such as insulin pumps have reported attacks where the attacker tries to modify the expected treatment. The attacker can copy the full elements of a remote control: Awaken the insulin pump, stop/continue the insulin infusion, or infuse a bolus dosage of insulin into the human body leading to the death of the victim.
Epilepsy and Parkinson’s disease patients using brain implants could be vulnerable to similar attacks. “Imagine what a terrorist could do with access to a politician’s mind or the repercussions if someone could alter how you act and think. These are scenarios that are unlikely to remain purely in the realm of science fiction for much longer,” says Heena Rathore, US-based data scientist, who works on wireless medical device security. “The future of neurological implants is bright, but even a single high-profile incident could irreparably damage public confidence on the safety of these devices, so the risk of brain jacking should be taken seriously before it’s too late.”
Implantable devices such as insulin pumps have reported attacks where the attacker tries to modify the expected treatment.
Taking inspiration from nature, Rathore is working to find solutions to these problems, especially those related to brain implants and insulin pumps. The latter injects insulin in the patients. It is monitored by an external computer or handheld computer that communicates with the implanted device wirelessly. Securing the communication between the insulin injector and the computer is essential to prevent access by unauthorised personnel. According to Rathore, deep learning and machine learning methods can be used to establish trust between the injector and the regulator, thus preventing any third-party from manipulating the system. “We can improve the operations of the implanted device and at the same time provide security against external threats. Taking a cue from the brain, we can mimic how it works and create a network of neurons to learn to distinguish between the trusted networks and the hostile networks; we can then control an external device mounted on the patient. Identification of these threats will be a great step to prevent an attack,” she adds.
Implantable medical devices range from neuro-stimulators for the brain, gastric stimulators for the stomach, cardiac defibrillators for the heart, a cochlear implant for hearing, to drug delivery system such as insulin for diabetic patients. Devices are now connected to each other and to the internet, which leads to the use of the term Internet of Medical Things (IoMT). To enable this, devices have WiFi/cellular chips on them so that they can talk to each other, in addition to the traditional roles of sensing and actuating.