Length of neuro-physiological signals of emotion compared to feeling time

Question

In "Determinants of Emotion Duration and Underlying Psychological and Neural Mechanisms" by Verduyn et al. discuss how certain physiological signals are outlasted by actual feelings:

In particular, whereas the duration of expressive (e.g., facial expression; [1]), and (neuro)physiological components (e.g., cardiovascular recovery; [2]) seems to rarely exceed a few minutes, feelings and associated cognitions may persist for hours or longer (e.g., [3], [4]).

Is this true for all external physiological signals including Galvanic Skin Response (GSR) and Electromyography (EMG)? I would find this surprising given that there is evidence of the central nervous system being related relating physical to social pain, as shown in "Acetaminophen Reduces Social Pain: Behavioral and Neural Evidence" Dewall et al., so I would expect some external evidence of this physical pain.

---

[1] Ekman, Paul, and Wallace V. Friesen. "Felt, false, and miserable smiles." Journal of nonverbal behavior 6.4 (1982): 238-252.

[2] L. Fredrickson, Barbara, and Robert W. Levenson. "Positive emotions speed recovery from the cardiovascular sequelae of negative emotions." Cognition & emotion 12.2 (1998): 191-220.

[3] Nolen-Hoeksema, Susan, and Jannay Morrow. "Effects of rumination and distraction on naturally occurring depressed mood." Cognition & Emotion 7.6 (1993): 561-570.

[4] Verduyn, Philippe, Iven Van Mechelen, and Francis Tuerlinckx. "The relation between event processing and the duration of emotional experience." Emotion 11.1 (2011): 20.

Answer 1

In "Simultaneous acquisition of corrugator electromyography and functional magnetic resonance imaging: A new method for objectively measuring affect and neural activity concurrently" Aaron et al. use face EMG to measure emotional reaction to photos. As shown in the figure below, the response only lasts a few seconds:

Getting a GSR response is more complicated due to the noise in the reading. In "Time-series analysis for rapid event-related skin conductance responses" by Bach et al., the authors figure out a mathematical approach for figuring out the event-related response. Both via a smoothed fitted gamma distribution and PCA component analysis are used below:

Both response measures seem to return to reference levels after 30 minutes.