Page 398 -

388 CHAPTER 13 Measuring the human

If you want to use physiological data to identify arousal, frustration, or other re-
sponses to specific interactions with a computer, you need to be able to synchronize
changes in physiology with user actions. Plainly speaking, if you know that the vari-
ability in a user's heart rate increased at a certain point in time, you won't be able to
interpret that change unless you know what the user was doing at the time. You are
likely to be keeping a textual log of user actions, tracking mouse movements, key
presses, and related information about the state of the application. Your physiological
data would similarly be recorded via software that would create fine-grained records
containing multiple measurements per second.
The first measurement challenge involves fine-grained measurements. Whereas
physiological data are essentially continuous, tracking of events on the computer
may not be. Fine-grained timing information may require using system clocks which
operate on the order of milliseconds. Recording the number of internal clock “ticks”
between events is one way to get high-resolution event data (Scheirer et al., 2002).
Due to processing or hardware requirements, physiological data might be captured
on one computer while tasks are completed on another. This arrangement presents
the challenge of managing a fairly complex experimental setup. Besides the two
computers (one for the application and one for data collection), you have sensors,
analog-to-digital converters for converting the physiological signals into a form suit-
able for storage on the computer, potentially modified input devices, and possibly
other equipment for audio and video recording (Figure 13.7).

The experimental setup
Toshiba Main monitor Power Mac
laptop 8500/180

Video
ProComp camera
unit
BVP GSR

Secondary Hacked
monitor mouse
(displays (with two
clock only) Subject cables)
EMG
FIGURE 13.7
A complex experimental setup: physiological signals are collected by the ProComp analog-
to-digital converter and stored on the laptop; the mouse is modified to simultaneously send
control events to the computer and pulses to the analog-to-digital converter.
From Scheirer, J., Fernandez, R., Klein, J., Picard, R.W., 2002. Frustrating the user on purpose:
a step toward building an affective computer. Interacting with Computers 14, 93–118. © Elsevier.