A short memoir by Kevin Donohue, co-founder of Signal Solutions, LLC

I was in my office one day back in 2002, busy wrapping-up a research project I had done on ultrasonic signal processing to discriminate between benign and malignant breast tissue, when I got a phone call from one of our junior faculty members.  It was Dan Lau.

He had been meeting with faculty from medicine and biology looking for potential research collaborations to develop his career.  Dan said there was a guy in biology, Bruce O’Hara, who was working on a system for classifying sleep and wake in mice non-invasively using a piezoelectric film sensor, and the signals looked a lot like the ones he had seen me talking about in my ultrasound work.  He invited me to come over and meet with Bruce.

Dan was an outstanding signal processing researcher, and certainly had the capabilities to pursue this work himself. I was impressed, however, that a junior faculty was willing to let this project go for a better fit and potential benefit of a more senior faculty member.

I immediately walked over to Bruce’s office and Dan introduced us.  Bruce presented some of the work he had done at Stanford and showed graphs of signals from the experimental system they had built.  Dan was right, the signals looked like the ones I had spent a decade characterizing and applying to the classification of benign and malignant tissue from ultrasonic echoes.

From our first meeting, it was apparent that Bruce had a lot of energy and liked to talk.  He had a contagious enthusiasm with extensive knowledge of sleep that bordered on entertaining.  Based on the signals he showed me, I knew I would be able to develop a much less complicated set of features and classifier to hopefully make this a useful tool for sleep researchers.

A bond was formed, and we began working as a team to get funding for this work.  Fueled by Bruce’s optimism and enthusiasm, and my obsession to process new data and extract information, we managed to get funding that next year, and a collaboration with the life science group at Oak Ridge Laboratories, where our first high-throughput system (32 mice simultaneously) was adopted and put to use.  Our company, Signal Solutions, was formed shortly after that the grant funding had ended in 2009.