Page 394 -

384 CHAPTER 13 Measuring the human

Commonly used techniques for measuring cardiovascular activity include BVP
monitoring and electrocardiography (EKG). BVP sensors worn on fingers measure
changes in reflect light associated with changes in blood volume in finger capil-
laries. These measurements can be used as indirect measures of anxiety and other
emotional responses such as that have been found to be correlated with blood.
Heart-rate variability information can also be inferred from BVP data (Scheirer
et al., 2002). Electrocardiography measures the electrical current that causes the
heart to pump. Using sensors placed on different places on the body, EKG can mea-
sure heart rate, the interval between heartbeats, and heart-rate variability (Mandryk
and Inkpen, 2004).

13.4.1.3 Respiration
Just as certain stimuli can make our hearts beat faster, changes in mood can affect
our breathing. Arousal may make us breathe faster and some emotions can cause
irregular breathing (Mandryk and Inkpen, 2004). Respiratory measures are strongly
linked to cardiovascular activity (Stern et al., 2001).
A relatively straightforward approach to measuring respiration involves track-
ing the expansion and contraction of the chest cavity. Sensors that can measure how
far and how rapidly the chest moves with each breath can be attached to the thorax
(Stern et al., 2001; Mandryk and Inkpen, 2004) and even integrated into clothing
(Brady et al., 2005).
13.4.1.4 Muscle tension
The contraction of muscles creates electrical signals that can be detected through
electrodes placed on the muscle of interest, a technique known as electromyography
(EMG). Measurements on the jaw can reveal tensions associated with a clenched jaw.
Sensors on eyebrows or cheeks can detect muscle movements associated with frowns
or smiles, respectively. Mildly positive emotions lead to lower EMG readings over the
eyebrow and mildly higher activity over the cheek, relative to mildly negative emo-
tions. Reactions to specific emotional moods including sadness, fear, and happiness
have been studied as well, with less clear results (Cacioppo et al., 2000). EMG has also
been used as an input modality: one project investigated the use of an EMG armband
as a means of unobtrusively controlling a digital media player (Costanza et al., 2007).

13.4.1.5 Brain activity
Numerous techniques for directly and indirectly measuring brain activity have been
developed. Brain-imaging techniques provide detailed displays, but expensive equip-
ment and required medical expertise have limited their use in HCI research. Indirect
measures that use changes in electrical signals on the head to measure brain activity
provide less detail, but they are significantly easier to work with.
Electroencephalography (EEG) involves the use of electrodes distributed across
the scalp to measure brain activity in the cerebral cortex. Typically, this involves
placing a cap containing 128–256 electrodes on a participant's scalp (Figure 13.5).
These electrodes are used to measure electrical activity in various locations,