The effect of adverse childhood experiences on chronic pain and major depression in adulthood: a systematic review and meta-analysis

Background Adverse childhood experiences have been linked to increased multimorbidity, with physical and mental health consequences throughout life. Chronic pain is often associated with mood disorders, such as major depressive disorder (MDD); both have been linked to adverse childhood experiences. It is unclear how the effect of adverse childhood experiences on neural processing impacts on vulnerability to chronic pain, MDD, or both, and whether there are shared mechanisms. We aimed to assess evidence for central neural changes associated with adverse childhood experiences in subjects with chronic pain, MDD, or both using systematic review and meta-analysis. Methods Electronic databases were systematically searched for neuroimaging studies of adverse childhood experiences, with chronic pain, MDD, or both. Two independent reviewers screened title, abstracts, and full text, and assessed quality. After extraction of neuroimaging data, activation likelihood estimate meta-analysis was performed to identify significant brain regions associated with these comorbidities. Results Forty-nine of 2414 studies were eligible, of which 43 investigated adverse childhood experiences and MDD and six investigated adverse childhood experiences and chronic pain. None investigated adverse childhood experiences, chronic pain, and MDD together. Functional and structural brain abnormalities were identified in the superior frontal, lingual gyrus, hippocampus, insula, putamen, superior temporal, inferior temporal gyrus, and anterior cerebellum in patients with MDD exposed to adverse childhood experiences. In addition, brain function abnormalities were identified for patients with MDD or chronic pain and exposure to adverse childhood experiences in the cingulate gyrus, inferior parietal lobule, and precuneus in task-based functional MRI studies. Conclusions We found that adverse childhood experiences exposure can result in different functional and structural brain alterations in adults with MDD or chronic pain compared with those without adverse childhood experiences. Systematic review protocol PROSPERO CRD42021233989.

The authors highlight the need for further research to improve understanding of the neural mechanisms underlying chronic pain, adverse childhood experiences, and depression.
Chronic pain is defined as pain that lasts longer than 3 months, or pain that persists beyond normal healing time. 1 The prevalence of chronic pain in the UK is 35e53% of the population, increasing with older age, and often associated with comorbid conditions such as depression. 2e5 A recent systematic review found that over half of patients with fibromyalgia had co-existing major depressive disorder (MDD), 6e8 although the nature of the link between pain and depression is not clear. 7 At a global level, chronic pain conditions are the leading cause of years lived with disability, 9 and, independent of pain, MDD has a prevalence of approximately 10%. 10e12 Adverse childhood experiences (adverse childhood experiences) are defined as repeated aversive experiences that represent a deviation from the expected environment and require adaptation. 13,14 Examples include physical, sexual, and emotional abuse, parental illness, criminality, violence, neglect, and poverty. 13e17 Chronic pain has been found to be related to adverse childhood experiences 18 ; for example exposure to adverse childhood experiences may be a risk factor for fibromyalgia in later life. 19 As with chronic pain, adverse childhood experiences have been proposed as a risk factor for developing mood disorders later in life. 20,21 Adverse childhood experiences are associated with elevated risk for multimorbidity with both mental and physical health problems. Although reasons for this are poorly understood, long-term immune system and neurobiological changes may play a role in altering vulnerability to chronic pain and MDD in adult life. 22,23 The quality of parental care, nutrition, cognitive stimulation, and socioeconomic status during early child development have been shown to affect brain morphology and functionality throughout the life course. 24 Neuroimaging is an important tool in advancing understanding of the neural correlates of chronic pain 25 and depression, and allows exploration of the impact of adverse childhood experiences, with the potential to identify common mechanisms. 24,26e28 There have been some neuroimaging studies of chronic pain and adverse childhood experiences identifying involvement of various brain regions, such as the insular cortex, basal ganglia, amygdala, hippocampal cortex, and prefrontal cortex 16,17,26 ; however these studies had not examined the comorbidity of adverse childhood experiences and chronic pain. Adverse childhood experiences and MDD neuroimaging studies share several common features, with reported abnormalities of the hippocampus, amygdala, anterior insular, and superior temporal gyrus, 29e31 but these studies had revealed the abnormalities without the comorbidity of adverse childhood experiences and MDD. A notable brain region strongly implicated as affected by adverse childhood experiences and depression in neuroimaging studies is the hippocampus; specifically, there is robust evidence for a smaller hippocampal volume. 30,32e38 There is limited understanding of how comorbid chronic pain and MDD are affected by adverse childhood experiences.
In this systematic review and meta-analysis, we aimed to synthesise evidence from relevant neuroimaging studies to investigate the impact of adverse childhood experiences on the neural correlates of chronic pain, and depression, both alone and in combination.

Literature
After prospective registration on PROSPERO (https://www.crd. york.ac.uk/prospero/display_record.php?RecordID¼233989; Registration number: CRD42021233989), a systematic literature search, data extraction, and meta-analysis were carried out. The search was conducted using the following electronic databases, MEDLINE (OVID), Embase (OVID), the Cochrane Central Registry of Controlled Trials (CENTRAL) (the Cochrane Library), PsycINFO (OVID), PubMed, Neurosynth, Sleuth, and Web of Science. An advanced search strategy was developed using a list of MeSH and keywords in combination or alone, such as chronic pain, neuropathic pain, MDD, depression, anxiety, neuroimaging, MRI, and early life adversity. Searches were customised for each database. The searches were restricted to English language, whereas no publication status, or date restrictions were imposed on the systematic search. Searches were imported into Rayyan for removal of duplicates, screening, and study selection (https://www.rayyan.ai/).

Eligibility criteria
Eligible studies were on humans, were randomised or nonrandomised, or observational (cross-sectional or longitudinal, cohort). To be eligible for review, studies had to include participants with chronic pain, depression with a history of adverse childhood experiences, or both, and have quantitative data and neuroimaging measures (i.e. structural and functional MRI), positron emission tomography (PET) or single-photon emission computed tomography (SPECT). Participants/study population of selected papers were adults aged 16 yr or older. Chronic pain was defined as pain that had been present for at least 3 months, and adverse childhood experiences were defined as repeated aversive experiences that represent a deviation from the expected environment and require adaptation and had been identified by a questionnaire such as the Childhood Trauma Questionnaire (CTQ). 39 Standard depression rating scales 40e42 were converted to a summary measure using a recognised conversion table in order to allow comparisons. 42 Studies were excluded if patients had acute pain, there was no diagnosis of depression, or depression severity was mild. Studies involving participants with brain injury, brain tumours, stroke, or neurodegenerative conditions were also excluded.

Study selection
To investigate the associations between chronic pain, mood disorder, and the effects of adverse childhood experiences on brain structure and function, neuroimaging studies identified through systematic searching were assessed for inclusion by two independent reviewers (GA, EL). Firstly, title and abstract screening was carried out, assessing against the predefined inclusion and exclusion criteria. Secondly, where it was not possible to determine eligibility from the title and abstract, full-text screening was performed. When discrepancies were raised, they were resolved through discussion, or, when they remained, by consultation with a third party (DS, LC). Cohen's kappa 43 was computed to assess the agreement between two reviewers 'include', 'maybe include', or 'exclude' decisions during the full-text screening.
When the selection process was completed, data were extracted according to our predefined protocol. Data extracted from each study included publication details (e.g. authors, year of publication, country where the study was carried out) and demographic characteristics of the sample such as age and sex. Moreover, the clinical diagnosis was extracted for MDD and chronic pain along with type and scale of measurements for adverse childhood experiences, MDD, and chronic pain. The neuroimaging method used was also extracted along with the modality, the patient groups, use of contrast (if taskrelated functional MRI [fMRI]) and the resulting significant coordinates (x, y, z coordinates of the brain regions).

Quality assessment
There is no standardised tool for assessing the risk of bias in MRI (structural and functional), to the best of our knowledge. Therefore, we modified a version of the NewcastleeOttawa scale, 44,45 which is used for assessing the quality of non-randomised studies, including case-control and cohort studies, in metaanalyses. The risk of bias tool included four categories, study selection aspects (adequate case definition: MDD, chronic pain, and adverse childhood experiences, representative of the cases, selection of controls and definition of controls), comparability (age, sex, and other variables), exposure (ascertain of exposure and drop-out rate), and statistical interference (uncorrected P-value threshold and correction for multiple testing at each voxel).

Activation likelihood estimate meta-analysis method
A coordinate-based meta-analysis approach, activation likelihood estimate (ALE), of extracted neuroimaging results 46e48 was used to analyse studies within BrainMap GingerALE 3.0.2 (http:// www.brainmap.org/ale). Talairach space coordinates were converted into Montreal Neurological Institute (MNI) coordinates using the 'icbm2tal' transform. 49 To take into account the spatial statistical uncertainty of the reported foci in the ALE method, foci were modelled as a 3D Gaussian probability distribution. 48 Considering between-subject and betweentemplate variance for each study, the full width half maximum (FWHM) of the Gaussian probability was computed. 46 FWHM of the modelled probability was computed through combining the activation probabilities of the reported foci, for each analysed study. 46 By combining the activation probabilities of the reported foci, modelled activation maps were computed for each analysed study. The union of these modelled activation maps as ALE scores were calculated within a grey matter on a voxel-by-voxel basis. ALE scores are tested against a null distribution of random spatial events to test for statistical significance. 46 Cluster-level family-wise error (I) has greater sensitivity and specificity and is less prone to type one error with regard to the convergence in comparison with voxel-wise thresholding. In particular, the foci convergence is achieved over the random clustering of the foci-noise via testing against the null hypothesis of random spatial association between experiments. 46 Thus, the I was used to guide the meta-analysis. Statistical significance was set at a corrected threshold of P<0.05 (using a threshold derived from the permutation of 1000 cluster-forming events assuming a voxel level of P<0.05). Computations involving the same group of subjects were pooled into one experiment to control for sample overlap. 48,50 A random-effects analysis was used (GingerALE 3.0.2) with incorporation of variable uncertainty based on subject size, 46 where the subject size is determined as the smallest number of participants in the contrast of the groups. 51 We performed a meta-analysis across all fMRI and voxel-based morphology paradigms contrasting participants who experienced adverse childhood experiences with participants who did not experience adverse childhood experiences. In addition, where there was a sufficient number of relevant studies, we aimed to perform separate exploratory metaanalyses on MDD (or chronic pain) participants with adverse childhood experiences, MDD participants without adverse childhood experiences, separately for fMRI and VBM and fMRI MDD with adverse childhood experiences compared with healthy controls. An assessment of heterogeneity is frequently used in non-neuroimaging studies; however, there is not yet any accepted method to calculate this for coordinate-based metaanalyses. Therefore, we were unable to assess this formally.

Results
After removing duplicates, 228 studies on mood disorder and adverse childhood experiences and 15 studies on chronic pain and adverse childhood experiences were identified, from title and abstract screening. After title, abstract, and e where appropriate e full-text screening, 43 studies were identified as meeting inclusion criteria for mood disorders and adverse childhood experiences, with six studies for chronic pain and adverse childhood experiences. 32e38,52e93 No studies were identified of adverse childhood experiences with comorbid mood disorder and chronic pain. Cohen's kappa was computed with good agreement between the two reviewers: kappa¼0.69 (95% confidence interval [CI], 0.57e0.82; P<0.0001) for full-text study selection. Of the 108 full-text studies screened, 43 met inclusion criteria: 26 were structural MRI studies, seven were resting-state fMRI studies, and 10 were task-based fMRI studies (Supplementary material). Most of these studies were cohort studies or case-control studies. Details of the included structural studies can be found in Supplementary Table S2, and task-based fMRI studies can be found in Supplementary Table S3. Summary tables of imaging findings for the structural and task-based fMRI can be found in Tables 1 and 2, respectively. Table 3 summarises identified resting state fMRI studies: it was not possible to carry out ALE meta-analysis because of inadequately reported coordinates in those studies. The risk of bias is summarised in the Supplementary material. The assessment of study quality showed no study at high risk, nine studies at an intermediate risk of bias, and 40 studies at low risk of bias.
Structural MRI experiments, voxel-based morphology were split based upon disease status: 1. MDD participants exposed to adverse childhood experiences were compared with MDD participants not exposed to adverse childhood experiences (five experiments; 13 foci; 277 participants), which revealed eight significant clusters in the anterior, occipital, temporal, frontal lobe, and putamen and insula (further details in Table 4). 2. MDD participants and healthy subjects exposed to adverse childhood experiences were compared with MDD participants and healthy subjects not exposed to adverse childhood experiences (11 experiments; 26 foci; 628 participants), which revealed one significant cluster in the occipital and anterior lobe (further details in Table 4). 3. Chronic pain participants exposed to adverse childhood experiences compared with chronic pain participants not exposed to adverse childhood experiences: no studies were identified.
Task-based fMRI experiments were split based upon disease status: 1. MDD participants exposed to adverse childhood experiences compared with healthy subjects (five experiments; 18 foci; 164 participants) which revealed six significant clusters in the parietal, limbic, frontal, temporal lobe and putamen/globus pallidus (further details in Table 4). 2. MDD participants exposed to adverse childhood experiences compared with MDD participants not exposed adverse childhood experiences (seven experiments; 28 foci; 270 participants), which revealed two significant clusters in the occipital and parietal lobes (further details in Table 4). 3. MDD, or chronic pain participants or healthy subjects exposed to adverse childhood experiences compared with those not exposed to adverse childhood experiences (10  55 MDDþELA>MDD-ELA MDD without ELA, compared with MDD with ELA: smaller hippocampal volumes found in males only. Gerritsen and colleagues 35 MDDþCM>MDD-CM CM: no change in hippocampal volume. MDD: was associated with smaller hippocampal volume. MDD was related to smaller hippocampal volume, in participants with CM. Monninger and colleagues 67 LS Reduce cortical thickness in the right medial orbitofrontal cortex and increased depressive symptoms were associated with chronic LS during infancy. Reduce cortical thickness in the left medial orbitofrontal cortex was associated with chronic LS during childhood and negatively correlated with the left caudal ACC and the left parahippocampal surface area. CT of the right transverse temporal lobe and the left entorhinal cortex volume was inversely related to LS during adolescence. Oshri and colleagues 69 adverse childhood experiences Participants with higher adverse childhood experiences scores had smaller right amygdala volumes and smaller centralmedial nuclei. Moreover, participants with higher adverse childhood experiences scores with increased anxiety, depressive symptoms and alcohol use showed smaller basolateral amygdala volume. Peng and colleagues 70 MDD-CN>MDDþCN Compared with patients without CN, patients with CN showed increased WM densities in bilateral sublobar extra-nucleus and right brainstem midbrain Salokangas and colleagues 72 RODþadverse childhood experiences >ROD-adverse childhood experiences ROD was not found to be associated with changes in the amygdalaehippocampus complex in adulthood. ROD patients with experiences of physical abuse in early life showed that are mediating the reduction of frontal lobe. Van Harmelen and colleagues 80

CEM>NO CEM
Compared with non-CEM participants, CEM participants: reduction in bilateral dorsal medial prefrontal cortex, independent of comorbid psychopathology Vythilingam and colleagues 81 MDDþCA>MDD-CA Compared with participants with MDD without CA and HC, participants with MDD with CA showed bilateral smaller hippocampal volumes. Yang and colleagues 85 MDDþCM>MDD   experiments; 37 foci; 296 participants), which revealed one significant cluster in the limbic and parietal lobe (further details in Table 4). 4. Chronic pain participants exposed to adverse childhood experiences compared with chronic pain participants not exposed to adverse childhood experiences (two experiments; nine foci; 23 participants) were not analysed because of the small group size.

Discussion
Adverse childhood experiences involving stress, injury, or diseases can change the brain at different levels, including epigenetic, 94 cell biological, and systems and network levels. 95,96 Using a systematic review and meta-analysis, we have explored adverse childhood experiences as a potential modulator for chronic pain and depression in adulthood. Significant structural alterations for patients with depression and adverse childhood experiences in the hippocampus, insula, and putamen were identified, with functional alteration in the precuneus. Investigation of the effects of adverse childhood experiences on chronic pain or depression revealed significant functional changes in the posterior cingulate cortex (PCC) and precuneus (Fig 1). We have shown that adverse childhood experiences have been linked with alterations in the function of the brain in people with chronic pain. Epigenetic and epidemiological studies have shown a relationship between adverse childhood experiences and chronic pain in later life 94,96e100 ; however, we have identified a gap in the neuroimaging literature studying the long-term impact of adverse childhood experiences on neural processing in people with chronic pain. There are neuroimaging studies examining psychosocial adversities (e.g. physical, sexual, emotional abuse), but there are no studies examining non-psychosocial adversities such as those arising from serious medical conditions (e.g. asthma, cancer, epilepsy in early life) which have often been overlooked. 101 The reward-motivation network (including the prefrontal cortex, nucleus accumbens, hippocampus, and ventral tegmentum) and the descending pain modulatory system (prefrontal cortex, anterior cingulate cortex [ACC], amygdala, hypothalamus) are implicated in vulnerability to painful conditions with evidence for structural, functional, and neurochemical alterations in the brain (Table 5). 102 Other areas that might not affect risk of developing chronic pain, but are relevant to pain perception include changes in the insula, thalamus, orbitofrontal, primary and secondary somatosensory cortex, and dorsal ACC extending to mid-ACC. 102,103 There is also an extensive literature around the involvement of the reward-motivation and salience networks in people with MDD. 104,105 The neural correlates of emotional processing have been well studied in people who have experienced adverse childhood experiences, although reward processing has received less attention, but there is evidence that suggests a deficit in reward sensitivity. 106,107 Our findings have demonstrated an impact of adverse childhood experiences on neuroanatomically plausible areas associated with mood and pain processing. Major cortical projections from the spinothalamic tracts (associated with pain) include the posterior insula, parietal operculum, and mid-cingulate cortex. 108 Posterior insula and parietal operculum stimulation and focal seizures can trigger pain, lesions can cause pain deficits, and cortical insula damage has been correlated with neuropathic pain. 108 Other regions, including    Effect of adverse childhood experiences on pain and depression in adulthood -737  anterior insula and anterior cingulate, influence the experience of pain but are also activated by other aversive experiences generating negative affect and anxiety. 108e110 Lesions in the posterior insula 111 and mid-cingulate cortex 112 can result in patients continuing to recognise pain but with reduced or absent suffering; pain asymbolia. 113 The anterior insula is thought to alter responsiveness to specific stimuli 114 because of its involvement in the salience network, 108 which mediates changes in the default mode network and frontoparietal network facilitating responses to salient stimuli. 115 Notably, a systematic review on childhood adversity and neural development highlighted the importance of the salience network, 116 with increased insula activation in children exposed to violence. 117e121 A meta-analysis of functional imaging studies reported that activation of the PCC was associated with the experience of pain and episodic memory retrieval 122 consistent with earlier reports. 123e125 The anterior mid-cingulate cortex (aMCC) is also strongly associated with the experience of pain, 108,126 and for some patients, small lesions of the aMCC relieve the distress of intractable pain. 112 Relative to the aMCC though, a larger and more posterior part of the cingulate is also associated with pain. 122 Furthermore, there is evidence for a rostralecaudal segregation of function, with the anterior PCC region more associated with pain and the posterior PCC region more associated with memory. 122 PCC activation appears consistently associated with the emotional salience of stimuli and, despite evidence for segregation, there is a tendency for emotion and memory-related activations to overlap. 122 Our meta-analysis study is consistent with these reports, such that patients with chronic pain and MDD were found to have reproducibly abnormal PCC activity. Furthermore, it is worth noting that healthy subjects with a history of adverse childhood experiences also have abnormal PCC activation. Consequently, abnormal activation of this region may be associated more with emotion and memory-related activations of the PCC rather than experiences of pain.
The hippocampus has an important role in the storage and retrieval of long-term explicit memories 127 and in associative learning. 128 It also has an important role in terminating the stress response via its regulatory role in the hypothalamicepituitaryeadrenal axis. 129 Hippocampal volume and functional abnormalities have been reported inconsistently in deprivation-exposed children. 116 Nevertheless, in children exposed to threat-related adversity, reduced hippocampal volumes have been reported, consistent with our findings including reduced activation during a memory task. 116 Reduced hippocampal volume has been reported in patients with chronic pain: fibromyalgia, 130 complex regional pain syndrome (CRPS), and chronic back pain (CBP), 131 with evidence for learning and memory deficits in patients with chronic pain. 132 Chronic pain and depression comorbidity is common, and it has been suggested that hippocampal abnormalities observed in patients with chronic pain may be related to the mood component. 133 Anhedonia, chronic pain, and depression have all been correlated with blunted striatum activation which may be related to dopaminergic abnormalities modulated by hippocampal afferents. 134 Studies on animals and humans have revealed reduced neurogenesis, neuroplasticity, neurotropic factors and increased hippocampal inflammation in both patients with depression and chronic pain. 132 The reward system includes the caudate and putamen, 135 with these regions responding to both the receipt and anticipation of reward. 136 Consistent with anhedonia being a prominent clinical feature of depression, the brain reward circuitry appears abnormal in depression with functional dysregulations and brain structure changes, 137 in accordance with our findings, such as blunted striatal responses to anticipated reward in maltreated children. 106,107 Blunted reward-linked striatal activity has also been observed in patients with adverse postnatal experiences. 138 Dysfunction of the basal ganglia during reward anticipation in humans exposed to childhood adversity may reflect abnormalities in dopaminergic circuits reported in animal studies. 139

Limitations
Potential limitations of this study should be noted. First, an assessment of heterogeneity is often done in non-neuroimaging studies, but there is no methodology for assessing heterogeneity for ALE or other coordinate-based meta-analyses. 141 The ALE method provides an estimation of the probability that an activity in a specific region may differ between groups of patients and not an estimate of the mean difference in the regional signal change. 142 Thus, traditional measures of heterogeneity are not applicable. 141,143,144 Second, the limited number of neuroimaging studies available in the literature investigating chronic pain with adverse childhood experiences limited the metaanalysis results. Third, adverse childhood experiences have been associated with other psychiatric disorders, such as posttraumatic stress disorder (PTSD) and anxiety, whereas our study focused only on patients with depression. Recall bias, 145 Table 5 Listing specific brain areas and their possible relation to pain, chronic pain, MDD and adverse childhood experiences. ACC, anterior cingulate cortex; adverse childhood experiences, adverse childhood experiences; MDD, major depressive disorder; PCC, posterior cingulate cortex.

Regions Involvement of MDD, chronic pain, and adverse childhood experiences
Prefrontal cortex Involvement in vulnerability to painful conditions (reward-motivation network and descending pain modulatory system). 102 Involvement of the reward-motivation and salience networks in people with MDD. 104,105 Deficit in reward sensitivity and reward anticipation in adverse childhood experiences. 106,107 ACC Individuals who are vulnerable to painful conditions may have changes in the reward-motivation and descending pain modulatory systems in the brain. 102 These changes may not necessarily affect the risk of developing chronic pain, but they do play a role in how pain is perceived. 103,108,126 Activated by other aversive experiences, leading to negative emotions and anxiety. 108e110 MDD may have alterations in the reward-motivation and salience networks. 104,105 adverse childhood experiences have been linked to a deficit in reward sensitivity and reward anticipation. 106,107 Amygdala Vulnerability to painful conditions may have changes in the descending pain modulatory system in the brain. 102 They play a role in how pain is perceived. 103 MDD have alterations in the reward-motivation and salience networks 104,105 which can contribute to their vulnerability to painful conditions. Hypothalamus/thalamus Vulnerability to painful conditions (descending pain modulatory system), 102 may not affect risk of developing chronic pain but is a relevant region to pain perception. 103 Involvement with the MDD. 104,105 Nucleus accumbens Individuals who are at risk of experiencing painful conditions may have changes in the reward-motivation network in the brain. 102 MDD may have alterations in the reward-motivation and salience networks, 104,105 which can make them more susceptible to painful conditions. adverse childhood experiences have been linked to a deficit in reward sensitivity. 106,107 Anhedonia, chronic pain, and depression is linked to dopamine abnormalities affected by inputs from the hippocampus. Research also suggests that childhood adversity can lead to dysfunction of the basal ganglia during reward anticipation. Ventral tegmentum Vulnerability to painful conditions (reward-motivation network). 102 Involvement of the reward-motivation and salience networks in people with MDD. 104,105 Deficit in reward sensitivity and dysfunction during reward anticipation in humans exposed to adverse childhood experiences. 106,107 Hippocampus Vulnerability to painful conditions (reward-motivation network), is a relevant region to pain perception. 103 Involvement of the reward-motivation and salience networks in people with MDD. 104,105 Children exposed to traumatic events have smaller hippocampus and weaker activity in that region during memory tasks. 116 This is associated with difficulties in learning and memory in individuals with chronic pain. 130e132 Changes in the structure of the hippocampus, which is often impacted by chronic pain, may contribute to the emergence of depression. 133,134 Insula It is an important area in regard to the perception of pain. 103,108 It is activated by other negative experiences as well, leading to negative emotions and anxiety. 108e110 Blunted signal in aversive stimuli in MDD. Mediates changes in the default mode network and frontoparietal network facilitating responses to salient stimuli. 108,115 Increased insula activation in children exposed to violence. 116e121 Orbitofrontal cortex It is still an area of the brain that plays a role in the perception of pain. 103 MDD may have alterations in the reward-motivation and salience networks. 104,105 Primary/secondary somatosensory cortex A significant area in terms of pain. 103

PCC
It is a significant area in terms of the sensation and perception of pain, 103 and episodic memory retrieval. 122 MDD have been found to have abnormal activity in this region, and healthy individuals who have experienced adverse childhood experiences. PCC activation may be more related to the emotional and memory-related aspects of stimuli, rather than the actual experience of pain. 122 unfortunately, cannot be eliminated because of the systematic error that occurs when the participants did not remember previous adverse childhood experiences accurately when answering the questionnaire. Prospective longitudinal cohort studies and pre-existing datasets that include information collected during childhood, which can be further linked to healthcare records in later life (e.g. Generation Scotland: STRADL-Pain, 146e148 ALSPAC 149 ), are approaches to address this limitation.

Conclusions
This review has investigated the neural correlates of adverse childhood experiences on chronic pain and depression. It is notable that there are few neuroimaging investigations into patients with chronic pain who have experienced adverse childhood experiences. Nevertheless, significant brain structure and function correlations with adverse childhood experiences were observed in people with major depressive disorder or chronic pain in comparison with healthy controls.
In the former, correlations were found in the hippocampus, superior insula, putamen, and precuneus, and in the latter, significant correlations were found in the dorsal anterior cingulate and precuneus. Our results indicate the existence of brain structural and functional abnormalities associated with adverse childhood experiences e some of which may be characteristic for adverse childhood experiences and others more related to comorbid depression and chronic pain. Moreover, they are suggesting a shared neural correlates for comorbidity and possibly increasing the vulnerability to develop later in life depression, chronic pain, or both.