A total of 28 healthy adults (ages 18 - 58; M = 31,61, SD = 11,50; 12 male, 16 female) were recruited via advertisement flyers. Written informed consent was obtained from all subjects prior to research. Subjects were seated in a relaxed position and attached to Galvanic Skin Response (GSR) sensor connected to Arduino computer that streamed recording data to personal computer at 200Hz sample rate. Two surface electrodes were attached to the index and the middle finger of the dominant hand. The room was kept quiet and at a comfortable temperature. Researcher observed GSR data on a personal computer and waited for subjects to achieve their baseline. Then researcher asked subjects to perform a voluntary inhalation and hold their breath for few seconds. Change in GSR was noted and afterwards researcher told the subject to relax and waited for baseline to be achieved again. Five consequent inhalations were performed by each subject and changes in GSR were noted accordingly. 28 subjects performed five voluntary inhalations each, resulting in 140 voluntary inhalations. Each subject’s mean and standard deviation of change in GSR during inhalation was calculated. Overall, GSR was reduced in all subjects (M = 34, 15%, SD = 5, 52%). In order to determine whether the drop in GSR was statistically significant t-test for paired samples was used (t = 8,25, p < 0,01). Our preliminary results suggest that inhalation and breath holding affect GSR in a way that GSR values drop, which may indicate increased anxiety or arousal. Such link between inhalation and anxiety may be observed in everyday life when one holds breath due to being exposed to stressful or arousing stimuli. However, it should be noted, there are some possible problems when using GSR as a method. For example, external factors (temperature, humidity, etc.) affect GSR measurments and galvanic skin responses are delayed 1-3 seconds, which can all lead to inconsistent results.

Pilot trial was published here (page 128)

We are planning a bigger trial in Slovenia in Summer 2023.