Neural correlates of thermal stimulation during active touch

Abstract

Introduction: Thermal feedback technologies have been explored in human-computer interaction to provide secondary information and enhance the overall user experience. Unlike fast-response haptic modalities such as vibration and force feedback, the human brain's processes associated with thermal feedback are not fully understood.

Methods: In this study, we utilize electroencephalography (EEG) brain imaging to systematically examine the neural correlates associated with a wide range of thermal stimuli, including 9, 15, 32, and 42°C, during active touch at the fingertip. A custom experimental setup is developed to provide thermal stimulation at the desirable temperature levels. A total of 30 participants are recruited to experience the four levels of thermal stimulation by actively touching a thermal stimulation unit with the index finger while recording brain activities via EEG. Time-frequency analysis and power spectral density (PSD) of the EEG data are utilized to analyze the delta, theta, alpha, beta, and gamma frequency bands.

Results: The results show that the delta, theta, and alpha PSDs of 9 and 15°C stimuli are significantly higher than the PSDs of 32 and 42°C in the right frontal area during the early stage of the stimulation, from 282 ms up to 1,108 ms (One-way ANOVA test, Holm-Bonferroni correction, p < 0.05). No significant differences in PSDs are found between 9 and 15°C thermal stimuli or between 32 and 42°C thermal stimuli.

Discussion: The findings of this study inform the development of thermal feedback system in human-computer interaction.

Keywords: EEG response; active touch; human-computer interaction; power spectral density; thermal sensation.

Figure 1

The experimental setup and thermal stimulation unit assembly. (A) Experimental setup. (B) An exploded rendering of the thermal stimulation unit assembly.

Figure 2

Schematic diagram of the experimental configuration.

Figure 3

The experimental protocol. (A) Experimental protocol with two types of sessions. (B) Experimental protocol of a trial.

Figure 4

Topographic plotting of delta, theta, and alpha power spectral density. The period from –3,000 ms to –2,000 ms is the baseline. The period from –2,000 to 0 ms (thermal onset) is the finger-reaching to touch the thermal pad. Thermal onset is the moment when the participant's fingertip contacts with the thermal pad. The period of 0 ms to 5,000 ms is the thermal stimulation. The red and blue colors indicate an increase and decrease in power respectively compared to the baseline respectively. (A) Distribution of the delta power spectral density. (B) Distribution of the theta power spectral density. (C) Distribution of the alpha power spectral density.

Figure 5

Spectrogram of the average PSD in the right frontal area (F6, F8, FC6, FT8, C6, and T8). The period from –3,000 ms to –2,000 ms is the baseline. The period from –2,000 to 0 ms (thermal onset) is the finger-reaching to touch the thermal pad. Thermal onset is the moment when the participant's fingertip contacts with the thermal pad. The period of 0 ms to 5,000 ms is the thermal stimulation. The yellow and blue colors indicate an increase and decrease in power respectively compared to the baseline respectively.

Figure 6

Average Delta power spectral density among 9°C, 15°C, 32°C, and 42°C thermal stimuli in the right frontal area (F6, F8, FC6, FT8, C6, and T8). (A) Time course delta power spectral density. One-way ANOVA test or Kruskal-Wallis test, Benjamini and Hochberg false discovery rate correction. Green vertical solid line, p < 0.01. (B) Mean delta power spectral density from 282 ms to 1,108 ms. One-way ANOVA test, Holm-Bonferroni correction,^*p < 0.05,^**p < 0.01,^***p < 0.001.

Figure 7

Average Theta power spectral density among 9°C, 15°C, 32°C, and 42°C thermal stimuli in the right frontal area (F6, F8, FC6, FT8, C6, and T8). (A) Time course theta power spectral density. One-way ANOVA or Kruskal-Wallis test, Benjamini, and Hochberg false discovery rate correction. Green vertical solid line, p < 0.01. (B) Mean theta power spectral density from 282 ms to 847 ms. One-way ANOVA test, Holm-Bonferroni correction,^**p < 0.01,^***p < 0.001.

Figure 8

Average Alpha power spectral density among 9°C, 15°C, 32°C, and 42°C thermal stimuli in the right frontal area (F6, F8, FC6, and FT8). (A) Time course alpha power spectral density. One-way ANOVA or Kruskal-Wallis test, Benjamini and Hochberg false discovery rate correction. Green vertical dashed line, p < 0.05. Green vertical solid line, p < 0.01. (B) Mean alpha power spectral density from 282 ms to 673 ms. One-way ANOVA test, Holm-Bonferroni correction, ^**p < 0.01.

Figure 9

Histograms of participants' perceived thermal sensation. VC, C, N, H, and VH indicate very cold, cold, neutral, hot, and very hot respectively. (A) 9°C thermal stimulation. (B) 15°C thermal stimulation. (C) 32°C thermal stimulation. (D) 42°C thermal stimulation.