In the 90's at MIT, I started new research on the measurement of emotion. Emotion changes with what matters most to people, and some professor's lab experiment is usually not at the very tippy top of that list.

To get data outside the lab, we created several new sensors. For example, we built electrodermal activity (EDA) sensors that could be worn on the palms (see the Galvactivator used for audience feedback as the first example). These sensors lit up and "glowed" when the audience was excited (live demos, live interaction, Q&A and laughter). The same sensors decayed exponentially in brightness, going "dim" when there was PowerPoint.

We built a wrist-worn version of the sensor that measured several things related to stress. We became particularly inspired by the requests of people on the autism spectrum, some of whom were non-speaking but typed to talk. They told us that their feelings were often misunderstood. Ironically, it is they who are usually accused of misunderstanding people's feelings! Could we build a sensor to help some of their feelings be "seen" and better understood? We invented versions that could work on the wrist and on the lower leg to be more comfortable (see picture below.)

We collected data indicating that people were becoming very stressed on the inside yet didn't look that way on the outside. People told us that sometimes they would get so overloaded that they would "shut down". Outwardly, they might tune out. Inwardly, they might be ready to explode. To tune-out was to calm down, but how could they signal that to others? Several told us, "If only I could be warned before I get to the point of a meltdown."

A lot of people are alexithymic - they don't fully read their internal feelings or know how to name them, or quantify them.

A technology designed to help measure the autonomic components of stress should be non-stressful to use. We made the sensors especially easy to pull on or off, and more.

Our ethic is that it's important that the wearer be in control of what is being sensed. After all, lack of control increases stress.

We obtained very surprising medical findings: For example, we can measure several kinds of seizures and their autonomic disruption, correlated with dangerous changes in brain activity.

We also obtained exciting findings related to pain measurement, sleep measurement and memory consolidation during sleep. More studies are underway today in many medical areas - also anxiety, depression, PTSD, bipolar, obesity, workplace stress, and more.

We called our MIT sensor the "iCalm" and built versions that both logged data and streamed it wirelessly. In 2010, Affectiva created and sold the "Q Sensor". This was the world's first comfortably wearable (24/7) and scientifically accurate electrodermal activity sensor, also sensing temperature and 3-axis accelerometer data and logging months of data at a time. (Electrodermal activity is measured in our case as skin conductance and is closely related to what used to be called "galvanic skin response (GSR)." It is also sometimes called Electrodermal Response (EDR)). We built it for our own research to provide quality data while being wearable long-term, and it was worn comfortably and safely by infants, children with disabilities, healthy adults and athletes, and elderly.

Affectiva was formed with two products, "Affdex", which was growing in the market research business, for ad pre-testing, and Q which was growing in medical. A little company has to pick one market to focus on, and our biggest investor was WPP, a world leader in advertising. Affectiva chose to focus on advertising and, in 2013, stopped selling sensors.

Picard then founded a new business to provide comfortably wearable clinical-quality data, Physiio International. In 2014, Physiio merged with Empatica Srl., an Italian affective computing company, to create a new business Empatica, Inc. Today, Empatica serves the growing demand for a beautiful, comfortable medical-quality wearable sensor and highest-quality analytics.