Informing participants impacts minute-by-minute blink count, but not average blink rate

Nina Shaafi Kabiri, Chris R. Brooks, Narayanan Kasthuri, Kevin C. Thomas, Peter J. Fried
Laboratory for Human Neurobiology
Department of Anatomy and Neurobiology
Boston University School of Medicine
650 Albany Street, Boston, MA, 02118


Objective. Counting eye blinks in humans can provide valuable physiological and behavioral data. However, awareness on the part of the subject that his or her eye blinks are being counted may impact blink rate and thus impugn its validity. The goal of this study was to provide direct experimental evaluation of this hypothesis.

Methods. Thirty adult males provided written consent prior to enrollment. All participants had normal vision and no known neurological disorders. The study conformed to the Declaration of Helsinki. All forms and procedures received prior approval by the Institutional Review Board at Boston University School of Medicine. Video of participants’ faces was recorded, from which blinks were counted. For each participant, there were a total of eight unique epochs, consisting of four situations (spontaneous, conversation, fixation, viewing images) and two conditions (naïve and informed). Four trained raters took part in counting; and treated only full occlusions of the eye as blinks. For each situation, minute-by-minute blink counts were entered into linear mixed model analysis with time and condition as repeated measures.

Results. Data from one participant was excluded for excessive blinking across all epochs. While viewing images, there was a main effect of time, F(2,252) = 10.561, p < 0.001, indicating blink counts increased overall with time. By contrast, there was no main effect of condition, F(1,28) = 0.131, p = 0.720, indicating that average blink rate did not vary according to whether the subject was informed that blinks were counted. There was a time x condition interaction between the two main factors, F(2,252) = 3.264, p = 0.001, indicating that blink rate over time was impacted by being informed that blinks were counted. Paired-sample t tests revealed that blink counts were lower in naïve vs. informed conditions during minute one (p = 0.009, uncorrected), but higher during minute three (p = 0.005, uncorrected), and were equivalent across all other time-points (all p values > 0.05, uncorrected). During fixation, there was a main effect of time, F(2,56) = 7.028, p = 0.002, indicating blink counts increased overall with time. However, neither the main effect of condition, nor its interaction with time was significant (all F values < 1, all p values > 0.4). No other effects were observed in any of the other situations, (all F values < 2.28, all p values > 0.05).

Conclusions. Being informed that ones blinks are being counted has a transitory impact on minute-to-minute blink counts, but not average blink rate while viewing images. Future studies are needed to determine if this applies to different blink situations.

Acknowledgements. This study was funded by the Neuroscience Research Unit of Pfizer, Inc. The authors declare no competing interests.