Patients with chronic pain lack somatic markers during decision-making

Figures & data

Figure 1 Decision-making structures important during the Iowa gambling task.

Notes: Trigger structures are the (A and C) V, and (B and C) A. The A mainly works as a trigger structure for the impulsive system, while the V is a trigger structure for the reflective system. Important supporting structures for the V are the dorsolateral prefrontal cortex (not shown) and (B and C) H. Effector structures include the hypothalamus, nucleus accumbens, and periaqueductal gray area, while sensory structures include the sensory brain stem nuclei such as the parabrachial nucleus and neurotransmitter nuclei. For effector structures and sensory structures, the figure only displays the (A and C) B. Important structures for processing information from the sensory structures are the (B and C) I, the (A) CC, (A) the P, as well as the somatosensory cortices (not shown). X and Z give the location of the two saggital and one transverse planes in the illustration.
Abbreviations: P, precuneus; B, brainstem; CC, cingulated cortex; V, ventromedial prefrontal cortex; H, hippocampus; A, amygdala; I, insula.

Figure 2 The generation of SCR during the IGT in the PCP and HC groups in the present study in light of the somatic marker hypothesis.

Notes: The schematic representation of the neuronal structures involved in generating and utilizing SCR during the IGT is from the somatic marker hypothesis.⁷ The top row shows the phases of the IGT. After choice n, the subject is presented with a reward and possibly a punishment (feedback phase). This visual feedback generates a feedback SCR. The subject ponders choice n + 1, which generates an anticipatory SCR (anticipatory phase). Anticipatory SCRs are interpreted by the brain and influence choice n + 1. The flow charts show the structures involved in generating SCRs and interpreting them during the IGT in HCs and in PCP. The arrows indicate the direction of information flow in a continuous process, initiated by the two trigger structures. The amygdala is more important as a trigger structure in the feedback phase due to its role in the impulsive system. The ventromedial prefrontal cortex is more important as a trigger structure in the anticipatory phase, and it triggers effector structures via the insula. The dotted lines indicate pathways that the current study suggests are abnormal in PCP, since they managed to generate normal feedback SCRs, but not normal anticipatory SCRs during pondering.
Abbreviations: SCR, skin conductance response; IGT, Iowa gambling task; PCP, patients with chronic pain; HC, healthy control; n, choice number; vmPFC, ventromedial prefrontal cortex.

Figure 3 The Iowa gambling task.

Notes: The Iowa gambling task is designed to test decision making. The figure illustrates the starting screen on a computerized version developed by the first author. The four decks from which subjects can choose have different and fixed rewards, while punishment frequencies vary between the decks, and punishment size also varies within the decks. From these characteristics, two decks are disadvantageous over time, while two decks are advantageous over time. Decision-making ability is scored by the number of cards the subject draws from the advantageous decks minus the cards drawn from the disadvantageous decks. The white text provides instructions to the subject in Norwegian (“You now have $2,000/press A, B, C, or D”). The green bar displays the amount of money the subject has, while the red bar displays the total sum of money the subject has borrowed to play. In this version, the placement of the four decks was randomized on the screen for each subject, and renamed A–D from left to right. In this article A–D refers to the decks by the classic nomenclature of the Iowa gambling task, not the letters displayed to the subjects.

Table 1 Demographic, clinical, and working memory measures in patients with chronic pain and matched healthy controls

	PCP	HCs	P	d
Education	4.67±2.4	5.26±2.7	0.484	0.23
Sex	18 (3 male)	19 (2 male)	0.597	0.18
Age	38.5±7.1	38.4±7.0	0.955	0.02
Pain NRS (prior 24 hour average)	6±1.64	1±1.29	0.000*	3.39
Pain NRS (at testing)	4.11±1.49	0.05±0.23	0.000*	3.86
Depression level	14.2±8.6	1.9±2.4	0.000*	1.96
Working memory measures
Letter–number sequencing	9.2±2.2	9.9±2.0	0.313	0.34
2-back score	47.9±3.8	50±2.9	0.059	0.64
PVSAT score	99.1±4.9	100.1±3.8	0.492	0.23

Notes: Depression level was determined with the Beck Depression Inventory II. Numbers are the numbers are the mean values ± standard deviation within groups of PCP with a pain self-rating of ≥4/10 for ≥6 months and in their healthy controls. Statistical differences between groups were explored with a two-tailed, two-sample Student’s t-test. Significance on the Student’s t-test is marked with * for P≤0.05. Effect size is calculated as Cohen’s d.

Abbreviations: PCP, patients with chronic pain; HCs, healthy controls; NRS, Numerical Rating Scale; PVSAT; Paced Visual Serial Addition Test.

Table 2 IGT total scores and switching in the learning phase (1–40 cards) and performance phase (41–100 cards), and brain volumes in PCP and matched HCs

	PCP	HCs	P	d
IGT score	5.2±28.6	5.7±31.0	0.959	0.02
IGT score, learning phase	−7.4±10.7	−9.2±13.3	0.665	0.14
IGT score, performance phase	−0.3±26.9	1.6±27.4	0.830	0.07
Switch total	0.35±0.15	0.26±0.15	0.080	0.59
Switch, learning phase	−6.13±0.16	−6.17±0.17	0.472	0.24
Switch, performance phase	−6.17±0.18	−6.29±0.16	0.031*	0.74
Brain volume, % of ICV
Total cortical gray matter	32.75±2.81	33.93±2.47	0.209	0.44
Hippocampus	0.53±0.04	0.54±0.05	0.872	0.06
Amygdala	0.25±0.02	0.25±0.02	0.556	0.20
Nucleus accumbens	0.07±0.01	0.08±0.01	0.034*	0.75
Brainstem	1.62±0.15	1.57±0.11	0.293	0.36

Notes: Numbers are the mean values ± standard deviation within groups of PCP with a pain self-rating of ≥4/10 for ≥6 months and in their healthy controls Cerebral volume is the combined volume of the two hemispheres in the % of ICV. Statistical differences between groups were explored with a two-tailed, two-sample Student’s t-test. Significance of the Student’s t-test is marked with * for P≤0.05. Effect size is calculated as Cohen’s d.

Abbreviations: IGT, Iowa gambling task; PCP, patients with chronic pain; HCs, healthy controls; ICV, intracranial volume.

Table 3 Correlations between pain level, the different IGT scores, and SCR before and during the IGT in PCP and matched HCs

	PCP					HCs
	Pain level	IGT score	IGT switching	SCR Before advantageous choices	SCR Before disadvantageous choices	Pain level	IGT score	IGT switching	SCR Before advantageous choices	SCR Before disadvantageous choices
Pain level	1*					1*
IGT score	−0.388	1*				−0.151	1*
IGT switching	0.262	−0.009	1*			0.172	−0.277	1*
SCR before advantageous choices	−0.02	0.103	0.091	1*		0.172	0.184	−0.119	1*
SCR before disadvantageous choices	0.044	0.063	0.203	0.539*	1*	0.215	0.568*	0.151	0.253	1*

Notes: Numbers are Spearman’s rho in PCP with a pain self-rating of ≥4/10 for ≥6 months and in their healthy controls. Statistical differences within groups were explored with a two-tailed Spearman’s rank–order correlation. Significance of the Student’s t-test is marked with * for P≤0.05.

Abbreviations: IGT, Iowa gambling task; SCR, skin conductance response; PCP, patients with chronic pain; HCs, healthy controls.

Figure 4 Autonomic measures from the anticipatory phase of the Iowa gambling task before drawing from either the advantageous or disadvantageous decks.

Notes: The autonomic measures were SCR (A), heart RR (B), and BP (C). The Y-axes denote the area under the respective measurement curves from 5.0 seconds before a card was picked from a deck. Points are split into PCP (red lines and squares) and their HCs (green lines and circles). The bars mark standard errors. •Significant within-group difference, P<0.05; ••significant between-group difference, P<0.05; *significant interactions between groups and card deck type, P<0.05.
Abbreviations: SCR, skin conductance response; RR, R wave to R wave intervals; BP, systolic blood pressure; PCP, patients with chronic pain; HCs, healthy controls.

Table 4 Results of the mixed ANOVA (group × choice type) analyses of the autonomic SCR, heart rate (RR), and BP before choosing from the disadvantageous or the advantageous decks in PCP and matched HCs

	F	P	ηp^2
SCR
Group	F(1,34)=2.869	0.099	0.078
Deck	F(1,34)=1.005	0.323	0.029
Group × deck	F(1,34)=6.195	0.018*	0.154
RR
Group	F(1,32)=1.917	0.176	0.057
Deck	F(1,32)=2.755	0.107	0.079
Group × deck	F(1,32)=0.542	0.467	0.017
BP
Group	F(1,32)=1.753	0.802	0.002
Deck	F(1,32)=1.916	0.176	0.056
Group × deck	F(1,32)=1.382	0.249	0.041
Simple main effects
SCR
HC group	F(1,18)=5.349	0.033*	0.229
PCP group	F(1,16)=1.363	0.260	0.079
Disadvantageous decks	F(1,34)=6.581	0.015*	0.162
Advantageous decks	F(1,34)=0.374	0.545	0.011

Notes: Relationship between group status and anticipatory autonomic activation among PCP with a pain self-rating of ≥4/10 for ≥6 months and in their healthy controls. The mixed ANOVA is by group (PCP or HCs) × deck type (advantageous or disadvantageous) for each of the three autonomic measures used in the study (SCR, RR, and BP). One patient was excluded from all SCR analyses due to technical problems with the recordings. Two controls were excluded from all BP and heart rate analyses because of excessive amount of extrasystoles. Measures are integrals of autonomic measures in the 5 seconds preceding the subject’s choice of card, which were either classified as advantageous or disadvantageous. Significance of the Student’s t-test is marked with * for P≥0.05.

Abbreviations: ANOVA, analysis of variance; SCR, skin conductance response; RR, R wave to R wave interval; BP, systolic blood pressure; PCP, patients with chronic pain; HCs, healthy controls.

Figure 5 SCR during the Iowa gambling task for pain patients and controls.

Notes: The relative changes in SCL for the 5 seconds prior to making advantageous choices (continuous line) or disadvantageous choices (dashed line) for PCP (red line) and HCs (green line), for illustrative purposes. Curves of significantly different SCRs are marked with * or §. The SCR was calculated as the area under a continuous SCL curve, with baseline as the SCL 5 seconds prior to making a choice. Simple main effects in a mixed design ANOVA showed that the SCR was significantly higher in HCs than in PCP for disadvantageous (§), but not advantageous choices. Only within the HC group was there a significant difference between the SCRs prior to making advantageous and disadvantageous choices (*).
Abbreviations: SCL, skin conductance level; SCR, skin conductance response; PCP, patients with chronic pain; HCs, healthy controls; ANOVA, analysis of variance.

Table 5 SCR before disadvantageous and advantageous choices, and after receiving reward and punishment cards in the IGT in PCP and matched HCs

	PCP	HCs	Between groups
			P	d
SCR before choice
Advantageous	−0.14±0.67	−0.02±0.49	0.545	0.2
Disadvantageous	−0.24±0.68	0.23±0.42	0.015*	0.86
Within groups
P	0.260	0.033*
d	0.16	0.55
SCR after choice
No punishment	0.24±0.79	0.08±0.50	0.467	0.25
Punishment	0.39±0.78	0.03±0.73	0.166	0.47
Within groups
P	0.224	0.742
d	0.19	0.08

Notes: Numbers are mean values within groups ± standard deviation in PCP with a pain self-rating of ≥4/10 for ≥6 months and in their healthy controls. Statistical differences between and within groups were explored with a two-tailed, two-sample Student’s t-test. Significance of the Student’s t-test is marked with * for P≤0.05. Effect size is calculated as Cohen’s d.

Abbreviations: SCR, skin conductance response; IGT, Iowa gambling task; PCP, patients with chronic pain; HCs, healthy controls.

Table 6 Cardiac autonomic regulation during the IGT and rest in PCP and matched HCs

	PCP		HCs		Between-groups tests
	Baseline	IGT	Baseline	IGT	Baseline		ΔIGT
					P	d	P	d
Blood pressure	126±18.6	123±13.8	125±17.1	123±12.7	0.617	0.17	0.892	0.05
Heart rate	73.5±8.4	75.1±13.6	72.4±8.1	75.2±13.6	0.466	0.25	0.161	0.48
Low-to-high frequency power ratio	2.41±1.51	2.66±3.49	1.65±1.16	2.55±3.39	0.310	0.35	0.135	0.52
Normalized low frequency power	62.1±16.1	61±13.8	54.8±14.6	57.8±17.1	0.583	0.19	0.294	0.36
Normalized high frequency power	33.2±13.99	35.22±12.9	41.86±14.18	38.46±16.4	0.516	0.22	0.145	0.50

Notes: Numbers are mean values within groups ± standard deviation in PCP with a pain self-rating of ≥4/10 for ≥6 months and in their healthy controls. Measures are calculated for the baseline period (resting period prior to task) and during the IGT. Between-groups tests were performed on the baseline data and on the group-averaged ΔIGT. Statistical differences between baseline and IGT within groups were explored with a two-tailed, two-sample Student’s t-test. Significance on the Student’s t-test is marked with * for P≤0.05. The effect size is calculated as Cohen’s d.

Abbreviations: IGT, Iowa gambling task; PCP, patients with chronic pain; HCs, healthy controls; ΔIGT, increase from baseline to IGT.

Table 7 Correlations between IGT behavior and autonomic measures and brain volumes in PCP and matched HCs

	PCP		HCs
	IGT score	Switching	IGT score	Switching
Autonomic activity before IGT choices
SCR advantageous	0.103	0.091	0.184	−0.119
SCR disadvantageous	0.063	0.203	0.568*	0.151
RR advantageous	0.06	0.289	0.083	−0.044
RR disadvantageous	−0.027	0.483	0.298	0.132
BP advantageous	−0.370	0.250	0.280	−0.402
BP disadvantageous	−0.194	0.417	−0.184	−0.338
Combined cerebral volume in % ICV
Total cortical gray matter	0.691*	−0.182	0.03	−0.042
Hippocampus	0.436	−0.319	0.129	−0.426
Amygdala	0.156	−0.701*	0.011	0.061
Nucleus accumbens	0.315	−0.152	0.108	−0.068
Brainstem	0.068	0.130	0.099	−0.355

Notes: Numbers are Spearman’s rho in PCP with a pain self-rating of ≥4/10 for ≥6 months and in their healthy controls. Cerebral volume is the combined volume of the two hemispheres in % of ICV. Statistical differences within groups were explored with a two-tailed Spearman’s rank–order correlation. Significance on Student’s t-test is marked with * for P≤0.05.

Abbreviations: IGT, Iowa gambling task; PCP, patients with chronic pain; HCs, healthy controls; SCR, skin conductance response; RR, R wave to R wave interval; BP, systolic blood pressure; ICV, intracranial volume.

ReimannMBecharaAThe somatic marker framework as a neurological theory of decision-making: review, conceptual comparisons, and future neuroeconomics researchJ Econ Psychol2010315767776

 Web of Science ®Google Scholar