IOO
delivered to the opposite hand). The false alarm rates were low for all auditory
stimulation conditions (see Table 1). An initial two-way ANOVA on the false alarm rates
was conducted, with auditory stimulus (left, right, none) and side of false alarm (left,
right) as the two within-subject factors. There was a significant main effect of auditory
stimulus that was mainly due to higher false alarm rates for trials with a sound as
compared to trials without any sound (F2,за = 5.45, p = .009). The main effect of side of
false alarm was not significant (p > .10). However, the interaction between auditory
stimulus and side of false alarm was significant (F2,38 = 5.63, p = .007). This interaction
was primarily due to subjects making more left-sided false alarms when the auditory
stimulus was on the left side and more right-sided false alarms when the auditory
stimulus was on the right side.
To further assess the nature of this interaction, we averaged the false alarm
rates across left and right sides and classified the false alarms as being same-sided when
a subject reported feeling something on the same side as the sound when no cutaneous
stimulus was provided or reported feeling something on same side as the sound even
though the cutaneous stimulus was delivered to the opposite hand. Similarly, opposite-
sided false alarms were computed by averaging across trials on which participants
reported feeling a somatosensory stimulus on a hand opposite a sound, regardless of
whether a somatosensory stimulus was delivered to the other hand or not. For the no
sound trials, the false alarm rates were averaged across the no somatosensory stimulus
and the somatosensory stimulus on the opposite hand trials. When a false alarm was
made on a trial in which a sound was delivered, participants were more likely to report