Galvanic Skin Response (aka Electrodermal Activity)
24. Dec 2008 
0 comments
I came across the Affective Diary the other day (Paper). There’s a youtube video but this blog entry sums it up nicely:
As you move through your day, you send and take pictures and send messages over your mobile. These messages and pictures are captured to the system which is connected to a body monitor that tells your pace, using a pedometer, and a galvanic skin response meter. When you’re excited, a little blob on the screen turns red and upright and when you’re relaxed the blob is blue and sleepy. The system allows you to watch a timeline of your blobs allowing you to see when and where you were most excited or agitated and even provides biofeedback.
One tester found that she was most agitated when her son was leaving to go back to Paris. By noting this, she learned she could tell her son she missed him and feel much better after he left instead of holding it in and getting herself upset. It’s a very humanist - and friendly - approach to technology.
So they’re correlating galvanic skin response with communication logs and using that information to reflect on experience — pretty cool. GSR is also used in lie detection tests and by the Church of Scientology to measure stress IIRC [not that that instutition is capable of lending this any credibility whatsoever]. I’m interested in playing around with this with my own kit so I’ve spent some time looking at building a GSR instrument.
Some Cornell students built a GSR monitor using a PIC, a whetstone bridge and an instrumentation amp. It’s interesting to see the effect of meditation has on the GSR plots: it turns the erratic signal into something quite recognizable. I’m critical about their bridge design as it seems to me that the single transducer would never produce a linear response. Their choice in resistor values must have put the transfer function in an approximately linear zone.
This paper uses an alternate approach that also measures the conductance response. Their aim is to detect drowsiness and it appears that the response is important for that. Their method uses a constant current injected across the skin with an instrumentation amp (an AD630 no less — used one of these in class) measuring the resulting voltage drop. It seems more robust to me.