News  General
Sense mood

In tune with your emotions

Smartphones will soon begin to act as intelligent companions with the skill to sense what our feelings are, and also have the capacity to advise us on what we need to do to enhance our lives.

Their sensors can gather data which advanced software can analyze, compare with millions of other samples in the cloud and derive conclusions from. In an instant, they can then give weighted probabilities, i.e. how likely they are. Examples already exist - the University of Rochester has a research project which can analyse stress levels in a voice recording already.

This man made brainpower won't simply be bound to smart phones, however. In the rapidly expanding world of IoT, any sensors can be used to gather these readings for further analysis in the cloud. Practical applications were recently shown at Internet World in London, where a radio player that can choose the next tune to play based on your mood was demonstrated.

Not only adapt, but anticipate

This idea has obvious medical implications. Mobilyze, from researchers at Northwestern University, acts at a virtual therapist which monitors activity over a period of several days before making its assessment. Apps which perform facial recognition types (smiling, frowning, etc) are already delivering promising results, too. Many non-pharmacological treatments are hard to access, including the 10% of Americans who experience a mood disorder each year. At the extreme end, the goal is to detect when there is an increased risk of suicide.

Psychologist David Mohr told WBBM Radio, Chicago:

We're trying to develop individual algorithms for each user that can determine specific states.

Another app is this field is called PRIORI, which detects mood swings via analysing voice pattern changes.

PRIORI 706x453

The project is called PRIORI because it is hoped it will yield a biological marker to prioritize bipolar disorder care to those who need it most urgently to stabilize their moods – especially in regions of the world with scarce mental health services. Often in such regions, the smartphone is the first - and only - contact people have with the internet.

desert tree stumps 700x250

Training improved the apps detection quality

Over time, richer data sets are gathered which facilitate more accurate analytics. Basic face gestures can already be determined with existing apps. These are bound to evolve, however, and could reach the point at which they are better at figuring out the persons mood than an actual human.

There are plenty of other "signals" a mood sensing app can analyse. GPS tracks location, and it has long been known that there is a correlation between the variety of locations a person visits and their depression level. It turns out staying in one place more is linked to a higher level. Sleep habits are also relevant. 

Prior to these innovations, subjects were asked to complete regular questionnaires, which quickly became tedious and provided unreliable results. The new way of constantly monitoring these signals hopes to eliminate this intrusive aspect which itself was skewing the results. With the explosion in IoT and wearable technology, this means the quality of the analysis can only improve over time.