Simultaneous measurement of electrodermal activity components correlated with age-related differences

Abstract

Electrodermal activity (EDA) measurements are influenced by various factors. Age-related psychological and physiological changes may be considered as one of the possible factors which may influence EDA measurements. In order to properly investigate the effects of such factors on EDA, techniques of precisely and simultaneously recording more than one EDA parameter are recommended. This study aims to explore the impact of age-related differences on EDA components through employing a new measuring technique, which is composed of a small front-end electronic box, DAQ card, and a laptop running LabVIEW software. It is dependent on the simultaneous recording of three EDA parameters: skin conductance (SC), skin potential (SP), and skin susceptance (SS) at the same skin site. EDA components as results of breathing, mathematical tasks, and image stimuli were recorded from 60 healthy participants simultaneously at the same skin site. They were categorized by age into young adults (ages 18–25), middle-aged adults (ages 30–40), and old adults (ages 50–70) years. It was found that skin potential responses (SPRs), and skin conductance level (SCL) (p < 0.001), were significantly decreased due to aging, but changes in other EDA parameters were nonsignificant (p > 0.05). Moreover, both tonic and phasic SS were the least affected and found to be more robust than SC and SP with respect to aging. The study suggests that it is important to take age into account in research studies where the mean aim of the study is to compare EDA parameters; however, in the meantime, the results from our small number and specific study population cannot be generalized to clinical applications.

Introduction

Electrodermal activity (EDA) is a set of physiological parameters of the sympathetic nervous system activity that originates from the eccrine sweat gland activity. In addition, it is one of the most widely used psychophysiological dependent variables [1] due to a strong link with the autonomic nervous system activity [2]. The investigation of EDA measurements began over 125 years ago with the pioneering studies of Fere in 1888 and Tarchanoff in 1889. Fere, through applying an external source of direct current observed changes in electrical resistance of the skin following emotional stimuli, and Tarchanoff observed changes in skin potential associated with emotional stimuli, which does not involve the use of any source of externally applied current [3].

EDA is mainly characterized by the two basic components: tonic (level) and phasic (response), each with various time scales and in associations with the stimuli. Tonic EDA is represented by the slow-changing baseline level, whereas phasic EDA is specified by a fast-changing component. Both tonic and phasic EDA components are generated under autonomic nerve control of the active organs of the skin [4], reflecting the evoked response of the sweat gland activity to external stimuli [3, 5].

The skin is the largest organ of the body. Stratum corneum is the outermost layer, which acts as a physical barrier. Changes in this layer have been associated with various factors such as aging, environmental conditions, and various cutaneous pathologies. Aging causes several physiological changes in the skin such as changes in its structural and biochemical properties as well as changes in neurosensory perception [6]. The first age-related changes in the adult skin appear between the third and fourth decade of life. In the fourth decade, a reduction in thickness and elasticity of the skin is likely to occur [3]. Also, insensible perspiration diminishes suddenly after the age of 60 years, perhaps as a result of decreasing skin blood flow [7]. Moreover, with aging, the number of active eccrine sweat glands and the sweat quantity per gland is reduced [8, 9]. Furthermore, skin water content is also reduced in aged skin, due to variations in the composition of amino acid, which may reduce the amount of the cutaneous natural moisturizing factor, thereby decreasing its capacity for water binding [6].

EDA recordings are possibly influenced by various external and internal factors. One of the physiological factors may be aging [3, 10]. The effects of age-related changes on EDA parameters are demonstrated in several studies [11–15]. However, the mechanism of age differences in EDA is unclear because several peripheral physiological components contribute to EDA [11]. Barontini et al. [12], for example, reported that skin conductance level (SCL) was significantly lower in older than in younger participants when tested across various age groups. Smaller skin conductance responses (SCRs) in older adults than younger adults during various stimuli were also reported by Shmavonian et al. [13] and Gavazzeni et al. [14]. In addition, Zelinski et al. [15] and Gavazzeni et al. [14] independently reported that older healthy individuals showed diminished SCRs amplitude than younger individuals during a memory task and in response to neutral and negatively tuned pictures, respectively. There is some inconsistency with respect to the influence of aging on EDA. Changes (i.e., reduction) in EDA due to aging were not observed in some other studies [16–19].

Changes in EDA due to age-related differences have previously been investigated in a number of studies [12–19]. However, they are carried out in different ways in which a single (either skin conductance (SC) or skin potential (SP) instead of focusing on simultaneously recording of all) EDA parameter is measured, which leads to inconsistency in the results of such studies. Moreover, in order to make direct comparisons between different EDA parameters and investigate the effects of age-related differences on these parameters, techniques for simultaneously recording EDA parameters at the same skin site are required [20]. Therefore, a required simultaneously recording system was developed and utilized in this study. The aim of this study was to recognize the impact of age-related differences on the tonic (level) and phasic (responses) electrodermal components through recording the three EDA (SC, SP, and skin susceptance (SS)) parameters simultaneously, by using a new, non-invasive bioimpedance technique.

Materials and methods

Study protocol and participants

Experiments were carried out on 60 healthy Caucasian volunteers in three age groups. Each group consisted of 20 participants in different age ranges as shown in Table 1. All participants were recruited from the University of Zakho and all of them gave written informed consent before taking part in the study.

Table 1.

Descriptive characteristics of the volunteers

Group	Female	Male	Age range (years)	Mean (years)
1	10	10	18–25	20.6 ± 2.03
2	10	10	30–40	33.8 ± 2.66
3	6	14	50–70	57.4 ± 6.70

The participants sat comfortably in a chair during the experiments in a sound-free room with the room temperature (22–23 °C). Before recording, 5 min was allowed for EDA electrodes stabilization after electrode application on one of the participants’ hands and then, SC, SP, and SS were simultaneously recorded.

In order to cause mental stress and observe possible specific differences in the level and in the habituation of autonomic responses, the participants were subjected to various external stimuli. The stimuli were:

1. Deep breath (breath): the participants were asked to take a deep breath for 4 s.

2. Cognitive (math): the participants were asked to subtract 17 successively from a starting number of 100 during 5 s, and

3. Vision (image): the participants were asked to look at a scary photo for 3 s.

There was a relaxation time for 60 s before and after each of the three stimuli to obtain a stable tonic level of the EDA measurements. Thus, the total recording session was 252 s for each participant. Any unnecessary speaking was not allowed for the participants and, at the relaxation time, they were asked to relax and to avoid bodily movement, particularly the testing hand during the whole session of data collection.

Instrumentation: skin admittance and potential measurement

A new custom-built measurement system for the recording of skin complex admittance and SP measurements simultaneously at the same electrode on the same skin site was utilized, consisting of a small front-end electronic box connected through a National Instruments DAQ card-NI USB-6211 to a personal laptop running custom-written software in LabVIEW, v. 14, similar to the system presented in Tronstad et al. [2] and Bari et al. [21–24]. A three-electrode measurement setup was employed, which consisted of one measuring electrode (M), one reference electrode (R), and a current-sink electrode (C) [2, 21–23]. The current-sink electrode, together with the reference electrode, served to provide a unipolar AC SC measurement below the measuring electrode. Simultaneously, the DC voltage between the measuring electrode and the reference electrode is used for SP recording. As in Tronstad et al. [2] and Bari et al. [21–24], a Howland current source was used. A 200-mV digital sine wave was produced by a PC, controlled through LabVIEW software, converted to analog by the DAQ card, and then fed to the Howland circuit. The Howland circuit in turn delivers a 20 Hz (less than 1 kHz as recommended [25]) AC of about 20 μA through the measuring electrode to the skin. The DAQ card receives the analog signals back from the skin via the front-end electronic box and converts to digital form. Then, the digitized signals are processed by differentiation in the PC LabVIEW and separated into a DC component for SP and an AC component for SC from the real part of the skin complex admittance signal and SS from quadrature part by phase-sensitive rectification. To detect variations in the reference site potential and to determine to what extent the reference site is electrodermally inactive, voltage sensing by analog-to-digital conversion at both terminals with software differencing was used. This is a key issue of concern for accurate recording of SP, as recommended by Boucsein [3].

Type of employed electrode and placement

The type of employed electrodes in this study was Kendall Kittycat 1050NPSM Ag/AgCl solid gel ECG neonatal electrodes with an active electrode area of 5.05 cm². Such electrodes were selected because they cause no or minimal wetting of the skin [25] and also have minimal effect on the state of hydration of the skin. In addition, they are better suited than wet gel electrodes for EDA measurements [25, 26] and hence are crucial for obtaining accurate results of EDA measurements. Electrodes were placed on the preferred arm of the participants. The M electrode was placed on the hypothenar site of the palm; the R electrode was placed on the apex of the elbow, which is an electrodermally inactive area and recommended by Fowles et al. [27]; and finally, the C electrode was placed on the underarm between M and R. Although abrasion by Edelberg [28] and skin drilling by Venables and Christie [29] were recommended as a pretreatment for the inactive site, no pretreatment of the skin was used in this study to avoid any risk of contamination.

Data and statistical analysis

In order to test whether the EDA signals (both tonic and phasic) differ among the three age groups, several parameters (Table 2) were calculated from the EDA measurements. In order to compute these parameters, the SCRs, SSRs, and SPRs onsets and peaks were first specified from responses due to breathing, the mathematics task, and the image stimulus for each subject through using the procedure presented in Bari et al. [21]. Also, the differences between electrodermal responses recorded from different age groups were evaluated statistically using the Statistical Package for Social Sciences (SPSS), IBM SPSS Statistics 22. The Kruskal-Wallis test was employed to assess the differences, followed by Dunn’s post hoc multiple pairwise comparison.

Table 2.

Overview of all EDA parameters extracted from the measurements and used in the statistical analysis

Parameter	Description	Unit
SPRET	Turning point of the SPR relative to the SCR peak	%
SCRs_Tris	Time from onset of SCR to peak SCR	s
SCRs_Amp	Amplitude of the skin conductance responses	μS
SCL	Skin conductance level	μS
SPRs_Amp	Amplitude of the skin potential responses	mV
SPL	Skin potential level	mV
SSRs_Amp	Amplitude of the skin susceptance responses	μS
SSL	Skin susceptance level	μS

SPRET (skin potential relative early turn) was extracted for all SC and SP responses for all participants to identify differences between the SCR and SPR waveforms. SPRET was calculated from the time of the SCR peak minus the time of the SPRs peak, divided by the time from SCRs onset to SCRs peak, and multiplied by 100%. This parameter was calculated to show the relative time difference between the turning points or peaks of the SPR and SCR waveforms [30].

Ethical approval

The protocol has been complied with all the relevant national regulations, institutional policies, and in accordance with the tenets of the Helsinki Declaration.

Results

Amplitude of EDA responses

Skin conductance responses

Figure 1 shows changes in median values of SCR amplitudes with respect to the different age groups and for the three stimuli. When data for SCRs amplitudes were analyzed, a nonsignificant (p > 0.05) difference between the age groups was observed as indicated by the Kruskal-Wallis test.

Fig. 1.

Box plot with medians, quartiles, and the minimum and maximum as whiskers shows amplitude of skin conductance responses for the three age groups with respect to breath, math, and image stimuli

Skin potential responses

SPR amplitudes changed (decreasing trend in SPRs) as results of different age groups and for all stimuli as seen in Fig. 2. These changes were statistically significant (p < 0.05) except for the mathematics task stimulus. Moreover, the group one was significantly different from three for both breath and image stimuli (Fig. 2) as indicated by Dunn’s post hoc pairwise multiple comparison tests.

Fig. 2.

Box plot with medians, quartiles, and the minimum and maximum as whiskers shows the amplitude of skin potential responses for the three age groups with respect to breath, math, and image stimuli

Skin susceptance responses

The results for SSR amplitudes with respect to the different age groups and for breath, math, and image stimuli are shown in Fig. 3. Response amplitudes (median values) were reduced as a function of age. However, once Kruskal-Wallis analysis was performed, no significant (p > 0.05) difference between the three groups was found.

Fig. 3.

Box plot with medians, quartiles, and the minimum and maximum as whiskers shows the amplitude of skin susceptance responses for the three age groups with respect to breath, math, and image stimuli

The summary of the statistical analysis on the amplitude of EDA responses is presented in Table 3, showing that only SPR amplitudes were statistically significantly different between the different age groups.

Table 3.

p values of the amplitude of EDA responses with respect to stimuli

Amplitude of EDA responses	Breath	Math	Image
	p	p	p
SCR_Amp	0.089	0.413	0.084
SPR_Amp	0.007	0.108	0.020
SSR_Amp	0.325	0.701	0.077

Levels of EDA

The tonic components (levels) of EDA are also impacted as results of various age groups. Figure 4a clearly reveals that the SCL is lowered with respect to age. The Kruskal-Wallis analysis also revealed a significant difference (p = 0.001) between groups. In addition, Dunn’s post hoc pairwise multiple comparison tests showed significant differences between groups one and three and groups two and three as indicated in Fig. 4a. Figure 4b shows that the SPL is negatively decreased as a result of aging. More specifically, participants with older ages have on average lower (decrease of negativity of the) SPL in contrast to the younger ones. However, once the Kruskal-Wallis test was performed on the magnitude of SPL, a nonsignificant (p = 0.133) difference between groups was obtained. The results for the mean value of SSL as a function of the aging are shown in Fig. 4c. SSL is changed with respect to aging, but, statistically, no significant (p = 0.20) differences were obtained with regard to age differences.

Fig. 4.

Box plot with medians, quartiles, and the minimum and maximum as whiskers, showing a decrease in the median value of a skin conductance level, b skin potential level, and c skin susceptance level according to age

Temporal components of the response

Skin potential relative early turn

Skin potential relative early turn (SPRET) as a timing parameter was calculated for all SCRs and SPRs from all test subjects for all stimuli. Figure 5 shows how changes in SPRET values coincide with aging for all stimuli. However, the Kruskal-Wallis analysis yielded a nonsignificant (p > 0.05) difference between data.

Fig. 5.

Box plot with medians, quartiles, and the min and max as whiskers, showing the SPRET percentage for the three age groups with respect to breath, math, and image stimuli

Rise time of SCRs

The timing parameter of the SCRs was also changed for different age groups (Fig. 6), but such changes were statistically nonsignificant (p > 0.05) as revealed by the Kruskal-Wallis test.

Fig. 6.

Box plot with medians, quartiles, and the min and max as whiskers, showing the rise time of skin conductance responses for the three age groups as with respect to breath, math, and image stimuli

The summary of the statistical analysis on the temporal components of the response is noted in Table 4, illustrating that all the changes in time-dependent EDA parameters between the different age groups for all stimuli are nonsignificant.

Table 4.

p values of the times of components of responses (SCRs) with respect to stimuli

Amplitude of EDA responses	Breath	Math	Image
	p	p	p
SPRET	0.658	0.988	0.499
SCRs_Tris	0.095	0.739	0.226

Discussion

In this study, influences of age-related differences on both tonic and phasic EDA components were assessed via employing a new EDA measurement system. Comparing with other traditional recording systems, which were used in previous studies, this system makes it possible to investigate the influence of age-related changes on both tonic (level) and phasic (response) EDA components at the same electrode on the same skin area. To the best of the authors’ knowledge, this is the first study to investigate the association between these three EDA parameters with respect to age-related differences and also the first study presenting SS to age-related changes.

The study results show that age-related differences have effects on EDA parameters; however, these effects on most EDA parameters were statistically nonsignificant, as p values were larger than 0.05 as shown in Tables 3 and 4. Figures 1, 2, and 3 reveal changes in phasic EDA (SCRs, SPRs, and SSRs) parameters with respect to age groups for breath, math, and image stimuli. As illustrated through the figures, aging leads to small changes in the amplitude of SCRs for all stimuli. However, these changes were statistically nonsignificant (p > 0.05). Moreover, a slight reduction in SCRs with respect to different age groups for image stimulus is seen in Fig. 3, which may be due to differences in emotional preference with age for negative stimuli (as image stimulus was a negative picture) [31]. These results are in agreement with [16–19] who could not demonstrate any influence of age on the SCRs. On the other hand, according to Andrew [32], reduction in amplitude of SCRs may be partly due to pronounced age-related variations in the hypothalamus, leading to further differences in the central triggering of autonomic nervous system reactions, which may interact with peripheral physiological differences between old and young individuals [28]. Figure 2 shows a decreasing trend in median values of the amplitude of SPRs as a result of age differences for all stimuli. In addition, the Kruskal-Wallis tests indicated a significant (p < 0.05) difference between groups with respect to breath and image stimuli. Moreover, when Dunn’s post hoc pairwise multiple comparison test was utilized, the group one was significantly different from the group three for both stimuli. These findings are in line with Garwood et al. [17], who reported that the aged individuals have a lower SPR amplitudes than young adults due to a decrease in sweat gland potential in old age, which is a result of the sweat gland ducts’ decay during aging. With regard to SSRs, no significant differences between median values of the amplitude of SSRs were obtained for the different age-related groups, although the amplitude of the SSRs was slightly reduced (Fig. 3). The results for the negative trend of SSR amplitudes are in agreement with our previous studies [21–23]. Unfortunately, to date, very few studies have been published on SSR recordings, and the authors could not find any literature, which investigated SSRs with respect to age-related changes for the purpose of comparing results.

Variations in tonic (level) EDA parameters as results of age differences can be seen in Fig. 4. In particular, SCL was significantly (p < 0.001) lowered with aging (Fig. 4a) which is attributing to greater skin resistance and reduced sweat gland activity in older participants. These results are in agreement with Barontini et al. [12], Catania et al. [16], Furchtgott and Busemeyer [18], and Ploufee and Stelmack [33] who found lower SCL in older participants. Edelberg [34] reported that the epidermal variations in aged skin would be too small to explain the detected increase in skin resistance. However, considerable changes in EDA properties appear during aging [34]. Also, in the psychological area, with age, there is a higher probability of depression [35]. It has been shown that individuals with depression exhibited lower SCL than healthy subjects [36]. SPL is also slightly influenced (Fig. 4b) with age-related changes, but the magnitude of the influence was statistically insignificant (p > 0.05) in agreement with Surwillo [37], who revealed that mean SPL was in the same range for both young and old groups. Small effects of age-related difference are also observed on SSL as shown in Fig. 4c, but these findings were not significant (p = 0.20). The reduction in SSL, in particular for the group three, is indicative of decreasing the electrical capacitance of the skin, which is proportional to the moisture content or hydration of the corneum [38].

Concerning the timing parameters, SPRET, which gives a more accurate interpretation of SPRs, was calculated for assessing the SP relative early turn as a function of age-related differences. Although on average the SPRET percentage was changed with respect to the age groups for all stimuli as seen in Fig. 5, these changes were statistically nonsignificant. As shown in Fig. 5, negative SPRET (i.e., SPRET < 0) is found with some responses, where the SPRs peaked later than the SCRs, whereas in all other responses, SPRs turn before SCRs (positive SPRET) in agreement with both Tronstad et al. [30] and Bari et al. [23]. The time between the onset and peak of the SCRs, which is known as the rise time (SCR_Tris), was also altered (Fig. 6) among the three age groups but nonsignificant (p > 0.05).

Conclusion

In conclusion, the effect of age-related differences on EDA components is investigated by using a new bioimpedance system, which is dependent on simultaneously recording of EDA by the same electrode and at the same skin site. Effects of age-related changes on EDA components (tonic and phasic) are studied by the addition of SS and SP measurements, which may facilitate the understanding of the complete mechanism of EDA due to aging changes. Our findings suggest that age-related changes may influence tonic (SCL) and phasic (SPRs) EDA, whereas such changes on other EDA parameters were not significant. Moreover, both tonic and phasic SS are least influenced and found to be more robust parameters than SC and SP with respect to age-related changes. Also, timing EDA (SPRET and SCR_Tris) parameters, which are associated with SCRs and SPRs, were not changed with age-related differences. This study implies that it is important to take age into account in research studies where the mean goal of the study is comparing EDA parameters; however, the observed age-related differences in this specific study population for the time being cannot be generalized to clinical applications.

Compliance with ethical standards

The protocol has been complied with all the relevant national regulations, institutional policies, and in accordance with the tenets of the Helsinki Declaration.

Conflict of interest

The authors declare that they have no conflict of interest.