Examining Changes in HRV and Emotion Following Artmaking with Three Different Art Materials

Summary

The goal of the protocol is to guide researchers in conducting experiments that are intended to measure changes in self-reported emotional response and heart rate variability following art making with different materials. The protocol can easily be adapted for use in a variety of behavioral conditions and activities.

Abstract

This protocol enables the examination of psychological and physiological responses to different types of behavioral engagements. Specifically, in this study example, the emotional response and changes in heart rate variability are examined in response to artmaking with three different art materials that vary in their levels of fluidity. This protocol can be adapted to examine other types of behavior or engagement in artmaking with other materials. There are several benefits to using this protocol. Firstly, the order randomization of the materials improves the probability that the response measured is associated with its qualities and not the order of presentation. Secondly, the continuous measuring of electrocardiogram enables the examination of changes in heart rate variability after engagement with each art material and changes that might occur during the artmaking itself. The advantages of this protocol should be considered with their limitations. The music listening is before each art making session; thus, the return to baseline can only be measured in the first two conditions. The return to baseline provides information on how fast individuals recover after response to working with each of the materials. Furthermore, a more liquid material instead of gouache paint with a brush, such as finger paints, provides more difference between materials. Finally, this protocol can be adapted to specific research needs.

Introduction

The purpose of this protocol is to examine the physiological and emotional response to artmaking in different conditions. In this case, the difference in response to art making with three different art materials varying in levels of fluidity is examined. The rationale behind the development of this experiment is to provide support for theories of art therapy claiming that artmaking with more fluid art materials is related to enhancing emotional expression¹. Heart rate variability (HRV) in general, and respiratory sinus arrythmia (RSA) specifically, is indicative of emotional engagement and regulation^2,3. In this study example, the order of the art materials used are randomized to control for an order effect. There are no other studies with this study design found.

The advantage of this method is that the ongoing measurement of heart rate variability (HRV) enables the examination of the physiological response to artmaking during the art making itself and it is noninvasive. This is in opposition to measuring bio-markers in blood or saliva following art making, which is important and relevant, but can be challenging to time accurately and requires uncomfortable (drooling in to a tube) or invasive (providing a blood sample) collection methods⁴.

This protocol can be adapted to measure response to a variety of behavioral activities, and art making with different materials. To do so, replace the artmaking with three materials with the desired behaviors to be examined. It is important to make sure that most elements of the behaviors being examined are similar other than the quality of the behavior being examined (i.e., the liquidity of the art material). Figure 1 is a flow chart of the experiment.

Protocol

This study was conducted with the approval of the Faculty of Social Welfare & Health Sciences at the University of Haifa, ethics committee. Informed consent was obtained, and research was performed in compliance with Helsinki guidelines for human welfare.

NOTE: The research is conducted at a table, using art materials and a portable, wireless and non-invasive electrocardiogram (ECG) recording device.

1. Participant selection

1. Recruit healthy participants between the ages of 18 and 40. For a wider range of ages, a division to sub-groups is recommended due to age related changes in heart rate variability.
2. Exclude participants with current or a history of heart disease as heart disease can affect heart rate variability, thus limiting the ability to examine whether change is due to the behavioral activity.
3. Invite all subjects to the laboratory between 08:00-15:00, to ensure similar conditions. Set any time frame as long as the experiment can be conducted within it.
4. Greet the participant and confirm that in the 2 hours before the experiment the participant did not drink coffee, eat, smoke or vigorously exercise.
5. Obtain written informed consent to ensure that participants know they are free to leave the experiment at any point in time without consequence and are aware of any risks and benefits to participation.

2. Experimental protocol

1. Adhere the biopatch two fingers below the sternum with 2 Ag-AgCl disposable electrodes. The ECG is sampled (1000 Hz, 16 bit) and recorded with the biopatch.
2. Examine that the values of the data collected (i.e., heart rate, HRV) appear on the graphic interface ensuring that data is being collected. The software will also transmit an audio alarm when the biopatch goes offline.
3. After the participant adapts to the environment and wears the biopatch, measure 5 min of resting heart rate in a sitting position to establish baseline HRV. Instruct the participant to sit still, and avoid chewing or talking, which affects the HRV.
4. Following the 5 min rest to establish baseline HRV, instruct participants to fill out pencil and paper questionnaires with demographic data, a question about art making habits and the self-assessment manikin (SAM)⁵. Instruct participants to mark the drawing that is closest to how they feel at the current moment (See Figure 2).
5. To establish a comparable emotional state, have participants listen to 5 min of relaxing music of their choice (nature sounds, new age or classical music) before each artmaking session. The relaxing music is purposely chosen by the participant to ensure that it is relaxing for them⁶.
6. Provide participants with a (50 cm x 35 cm) sheet of paper and instruct them to engage in a 10 min artmaking session (pencil, oil pastels or gouache paint). To reduce an order affect, randomize the order of the art materials.
   NOTE: Here, after drawing with pencil, the participants used oil pastels and then gouache paint.
   1. Encourage participants to engage in an exploration of the materials in any way they feel fit and use the entire 10 min for artmaking. Provide encouragement by reminding participants that the art product and its aesthetics are not important and that they can draw anything at all.
   2. Present the pencil with a pencil sharpener and an eraser.
   3. Instruct participants to report their emotional state using the SAM.
7. Turn on music and instruct participants to listen to 5 min of relaxing music of their choice (nature sounds, new age or classical music).
8. Place an open box of 12 colors of oil pastels on the table and instruct participants to use them in any way they choose including peeling the wrapper and breaking the pastel if needed.
   1. Replace the box of pastels for the next subject if their general appearance is significantly affected by use.
   2. Instruct participants to report their emotional state using the SAM.
9. Turn on music and instruct participants to listen to 5 min of relaxing music of their choice (nature sounds, new age or classical music).
10. Present gouache paint in primary colors (yellow, red and blue), black and white in a plastic palette, divided into 6 bowls and 4 mixing surfaces, to control the quantities of paint. In addition, provide a jar of water, a soft, flat-headed brush (size 6) and a cloth.
    1. Tell the subjects that they can request additional paint if needed.
    2. Provide a short explanation of mixing colors and cleaning the brushes.
    3. Instruct participants to report their emotional state using the SAM.
11. Remove the biopatch from the participant at the end of the session. Download the data recorded with the biopatch to the computer using the biopatch download log.

3. Data analysis

1. Convert the data to a text file that is appropriate for import to the visual analysis software (e.g., QRSTool)⁷, saving only the time stamp and voltage vectors, using a MATLAB script. The MATLAB script* is helpful in saving the timestamp and voltage vectors, while erasing the rest. This can be done manually but is very time consuming. The resulting text file example is presented in Figure 3.
2. Import the text file created by the MATLAB script to the QRSTool using the graphical user interface: File | Import | (Choose File). Make sure that the sample rate 1000 Hz button is pressed. The ECG will then be graphically displayed in the window of QRSTool.
3. Visually inspect the ECG series for artefacts. Identify the R component of the ECG using the QRS-tool software (See Figure 4). The software automatically detects the R spikes, by flagging all points with a red hourglass, above a threshold marked manually.
   NOTE: For more information contact the third author at: jallen@email.arizona.edu
   1. Visually inspect each ECG series alongside the identified heartbeats, time points and artefacts (missed or wrongfully identified heartbeats), which may have an excessive or missing heartbeat in the inter-beat intervals graph. Correct this manually by clicking to remove the extra heartbeat or adding one where missing.
   2. In areas of data with noise, not showing an identifiable QRS waveform, divide the data into several chunks, of 30 seconds at least.
      NOTE: Remember, tails of cropped data cannot be added together. Analyze these separately and create a mean of all chunks.
   3. Mark or divide the data according to various phases of the experiment. To do so, use the visual interface to cut off and save only the data for each event and export them separately to CMetX (described below). Divide the data based on the experiment stages described above (baseline, music listening and art making with the various materials).
4. When all data is clean, export to CMETX by using the visual interface File | Export, or directly by a command line. This will open the CMetX viewer window with an added line with the indices of HRV.
5. Export HRV indices to a CSV file for data analysis
   1. Use two indices of HRV: the estimated respiratory sinus arrythmia (RSA) as a vagal or parasympathetic index and the cardiac sympathetic index (CSI) – based on a Lorenz plot that estimates sympathetic influence. Figure 4 is an example of the output provided by the CMetX software.
      NOTE: Choosing the appropriate HRV index to represent vagal or sympathetic cardiac regulation can be tricky and requires some methodologically and theoretically based decisions that are beyond the scope of this presentation. For further information we recommend reading: Allen, J. J., Chambers, A. S., & Towers, D. N. (2007). The many metrics of cardiac chronotropy: A pragmatic primer and a brief comparison of metrics⁸.
6. Estimate HRV reactivity by subtracting the mean HRV index during music listening from the mean HRV index during the following artmaking session (i.e., positive values represent higher HRV during artmaking compared to the proceeding relaxation).

Representative Results

As mentioned above, the CMetX software provides several matrices of HRV. Here, the RSA measure, was calculated for each participant per each experimental condition. Figure 5 depicts mean RSA values (±SE) for the three artmaking tasks and their preceding resting periods. A within subjects ANOVA revealed a statistically significant main effect for artmaking (F^(1,49) = 26.155, p < .001), with a large effect size (η_p² = .348), which can be intuitively interpreted as thirty five percent of the variability of RSA explained by artmaking (compared to resting). A statistically significant interaction (F^(2,98) = 5.965, p = .004, η_p² = .109) indicated that the change in mean RSA levels was dependent on the art material. Pairwise comparisons (with a Sidak adjustment for multiple comparisons) demonstrated a significant change in mean RSA during artmaking with oil pastels (t₍₄₉₎ = 5.51, p < .0005, Cohen's D = 0.475) or gouache (t₍₄₉₎ = 3.63, p = .001, Cohen's D = 0.195) but not during drawing with pencil (t₍₄₉₎ = 1.40, p = .168, Cohen's D = 0.105).

Figure 5: RSA at rest and during artmaking. Please click here to view a larger version of this figure.

In addition to heart rate variability, the three self-reported emotional response measures of the SAM⁵ visual analogue scale were used to estimate whether there were differences in the emotional responses to the various art materials. Mean (±SD) values of the valence, arousal and dominance self-report measures following each artmaking task, as well as following a baseline resting period, on a centered scale ranging from -4 to 4 are detailed in Table 1. A repeated within subjects ANOVA demonstrated a statistically significant effect for art material on valence but not on arousal or dominance (Table 1). Post hoc pairwise comparisons revealed that, on average, emotional valence was more positive following painting with gouache in comparison to emotional valence following drawing with a pencil (p = 0.038), but not in comparison to baseline valence (p = 0.744).

	Baseline	Pencil	Oil Pastel	Gouache	F(3,147)	ηp2
	M (SD)	M (SD)	M (SD)	M (SD)
Valence	2.14 (±1.07)	1.96 (±1.65)	2.30 (±1.54)	2.60 (±1.55)	2.93*	0.056
Arousal	-1.72 (±1.84)	-1.86 (±2.13)	-1.90 (±1.88)	-1.66 (±2.11)	0.47	0.009
Dominance	-0.02 (±1.41)	0.16 (±1.54)	0.14 (±1.77)	0.16 (±1.82)	0.28	0.006
*p < 0.05

Table 1: Valence, arousal, and dominance measures of emotional response during rest and art-making.

For detailed results, please see: Emotional response and changes in heart rate variability following artmaking with three different art materials⁷.

This protocol could be easily used to examine differences in emotional response after and changes in HRV during and after a variety of activities, and do not have to be limited to artmaking. Furthermore, this protocol can be modified to examine differences in emotional and physiological response to artmaking with a wider variety of art materials as well.

Figure 1: A flow chart of the experiment, originally published in a previous publication and is printed here with permission⁷. Please click here to view a larger version of this figure.

Figure 2: An example of the SAM measure⁵. Please click here to view a larger version of this figure.

Figure 3: An example of what a text file prepared to be imported to QRSTool looks like. Please click here to view a larger version of this figure.

Figure 4: An example of the data exported by CMetX. Please click here to view a larger version of this figure.

Discussion

This protocol is designed to measure differences in RSA during artmaking with art materials varying in their levels of fluidity⁷. The biopatch^9,10 is a small round physiological monitoring telemetry device that attaches to a holster that connects to two Ag-AgCl disposable electrodes and enables the collection of physiological data, including basic ECG. The device stores and transmits vital sign data including ECG, heart rate, respiration rate, body orientation and activity. There are several critical steps in the protocol. It is very important to conduct some sort of returning to baseline activity. We chose listening to relaxing music, and let participants choose one of three types of music to ensure that indeed this music is relaxing for them. We also had the participants listen to the same music before engaging in art making with the goal of "neutralizing" any stimulation that participants may have entered the laboratory with that day. There are some disadvantages to allowing participants to choose the type of music they listen to as it could impact variability between how participants respond. However, the purpose of the music is to provide an opportunity to return to baseline and we believe that having participants choose how to return to baseline is preferred over the potential that the music would be annoying and thus missing its purpose of return to baseline.

Paying attention to ensure that the time one activity ended and the next one began is marked accurately is important. The absolute timestamps provided by all research tools should be synchronized to ensure that the ECG recordings are correctly divided according to the actual activities timing.

The SAM is a visual analogue scale measuring the valence, arousal, and dominance of emotional response administered here on paper and pencil form with 3 rows of 5 drawings of human figures, and a scale of 1-9 below each row (see Figure 2). The valence scale ranges from unhappy or sad to happy or joyful. The arousal scale ranges from calm or bored to stimulated or excited. The dominance scale ranges from submissive or "sense of being without control" to dominant or "in control". The first row is a valence scale, which has drawings that have a range of facial expressions starting with a smile and slowly transitioning to a frown. The smile indicates a very happy mood, while a frown indicates a very sad mood. There are 3 drawings in the middle, with a small smile, a neutral expression and a slight frown. The second-row measures arousal, and the human figure drawn has a large explosion like drawing in the middle of the human figure, which transitions to become smaller and smaller until the figure on the far right only has a dot in the mid-section, indicating a low level of arousal. The dominance scale at the bottom has a small figure in the center of the square that progressively gets larger, until the far-right figure emerges from the bounds of the square.

Since ECG artefacts, which may be caused by participant movements, or temporary disconnection of the biopatch, might erroneously identified as R waves, it may result in R-R (i.e., Inter-Beat-Intervals - IBI) and HRV miscalculations. In areas of data with noise (not showing an identifiable QRS wave), use QRSTool to divide the data into several long enough chunks, of 30 seconds at least. The visual interface of QRSTool enables the user to mark a certain section of data, if suspected to be noise and then crop it out, so that it is not included in the outputted data. Another advantage of dividing the data into the 30 second segments is that the metric of HRV will not be influenced by length of usable data, as longer recordings can produce higher HRV estimates.

CMetX software is a command-line based utility that calculates various metrics of HRV given a simple IBI series as an input. The following indexes are calculated: mean interbeat interval, mean heart rate, the average of the rate-transformed IBIs, standard deviation of IBIs, root mean square of differences between IBIs, mean of absolute value of consecutive IBI difference, proportion of consecutive IBI differences greater than 50 ms, Toichi Cardiac Vagal Index, Toichi cardiac sympathetic index, natural log of variance of IBI series, natural log of variance of filtered (.12-.40 Hz) IBI series, number of IBIs on which the metrics are based, which allows for loss for implementing the filter to band-limit RSA. All values are calculated for the IBIs that are retained after the filter is applied to band-limit the signal to calculate RSA. The filter results in a loss of 12 seconds of data at the beginning and 12 seconds at the end of the data segment. Further documentation and training videos on scoring ECG can be found on: https://jallen.faculty.arizona.edu/content/resources-and-downloads.

As mentioned earlier, the protocol can be modified to examine differences in emotional response and physiology to different types of behaviors. In addition, the protocol can be modified to use additional self-report measures as well. Parts of the ECG recordings from the beginning and end of sections may need to be removed if time synchronization is not optimal. The protocol is also not limited to 3 different types of activities or materials and can be expanded or restricted. Having said that, the protocol is limited to behaviors and activities that can take place in one room in a similar setting for all participants and all activities for each activity. This is because the environment may affect HRV and to ensure that the differences observed are due to the activity or behavior and not the environment.

The level of experience in art making, individuals' level of anxiety from art making, along with any other stressful experiences that may occur before the experimental session, may have an effect on individual's response. In this study excerpt, we asked participants regarding their experience in artmaking, however we did not find any difference based on artmaking experience (novice, hobby or professional). Further studies may want to take the other considerations mentioned in to account. Another suggestion would be to have open ended questions to probe individuals about the reasoning behind their emotional state.

In the future, this protocol can be used to compare response to drawing with regular art materials or digital media. Given the finding regarding oil pastels, we suggest conducting a further study to test the variability between a waterier type of paint and comparing finger painting with different types of paint brushes. The results of studies employing this protocol can be used to expand the body of knowledge regarding the effect of the use of different materials which in turn may be used to tailor art-based interventions to the specific needs of art therapy clients.

Materials

Name	Company	Catalog Number	Comments
Disposable Ag/AgCl electrodes	Biopac	EL501
Drawing paper	Stenoplast
Echo gateway	Medtronic	9600.0303
Eraser	Factis
Gouache paint	Giotto
HB pencil	Milan
Omnisense 3.9.7	Medtronic	9700.0269	Computer software
Oil pastels 12 colors	Talens
Zephyr biopatch	Medtronic	9600.041