Table 1.
'Overview of Included Functional Magnetic Resonance Imaging Studies Investigating the Association Between Childhood Maltreatment and Alterations in Resting-State Functional Connectivity in Children and Young People.
| Study | Sample Size1 | Mean Age (Age Range)1 | Gender (Female) | Ethnicity | Maltreatment Assessment | Analytic Method | Concurrent Psychopathology | Connectivity Findings2 | Other Findings |
|---|---|---|---|---|---|---|---|---|---|
| Cheng et al. (2021) | 223 (127 MT, 96 NMT) |
16 (11–19) |
61% | n/a | IR and/or SR CAT3 |
SBC | 44% in MT group, 32% in NMT group. | ↑ AMG - INS (AI) ↑ AMG - dACC ↑ AMG - HPC ↑ AMG - PCC/PCu ↑ AMG - OFC/vmPFC ↑ AMG - lPFC/dlPFC ↑ AMG - PoCG/PrCG ↑ AMG - TPJ/STG ↑ AMG - ITG |
n/a |
| Fareri et al. (2017) | 88 (41 MT, 47 NMT) |
12 (6–18) |
63% | n/a | PI CAT |
SBC | Participants above clinical cut-off were excluded. | ↑ VS – mPFC (BA9, BA10, BA24, BA32, and BA4)4 | Stronger ventral striatum–mPFC rsFC was associated with parent reports of more social problems in the PI group |
| Goetschius et al. (2020) | 175 | 16 (15–17) |
56% | 73% B 15% W 12% O |
PR CON |
GBNA | n/a |
↓ dACC – INS (AI) ↓ INS (AI) - AG |
Violence exposure was associate with higher rsFC heterogeneity (few shared connections) and lower network density (sparsity). |
| Herringa et al. (2013) | 64 | 18 | 47% | 95% W 3% B 2% O |
SR CON |
SBC | n/a | ↓ AMG - vACC/sgACC ↑ AMG - dlPFC ↓ HPC - vACC/sgACC |
Reduced AMG - vACC/sgACC and HPC - vACC/sgACC rsFC mediated (cross-sectionally) the association between maltreatment exposure severity and internalizing symptoms. |
| Herzberg et al. (2021) | 83 (41 MT, 42 NMT) | 13 (12–14) |
65% | n/a5 | PI CAT |
SBC and GBNA | Participants above clinical cut-off were excluded. | ↑ AMG - vACC/sgACC | Uncorrected exploratory analyses revealed that PI children showed reduced rsFC within the dorsal attention network (FEF and IPS). |
| Hoffmann et al. (2018) | 44 (19MT, 25CT) |
15 (12–16) |
54% | 62% W | IR CAT |
SBC | Participants did not meet clinical cut-off criteria. MT group showed greater symptoms than NMT. | ↓ vACC/sgACC - dlPFC ↓ vACC/sgACC - SMG ↓ vACC/sgACC - CUN |
Reduced sgACC–CUN rsFC mediated (cross-sectionally) the association between maltreatment exposure and higher depressive symptoms. |
| Marusak et al. (2015) | 33 (14 MT, 19 NMT) |
12 | 76% | 42% B 33% W 6% H 6% O 12% n/a |
SR CAT |
SBC | MT and NMT did not differ in symptoms severity. Cut-off clinical criteria n/a. | ↓ dACC - (AI + dACC) ↑ AMG - (AI + dACC) ↑ MI - (AI + dACC) ↑ dACC - (mPFC + PCC + AG) |
Salience network (i.e. AI+dACC covariance) to insula rsFC mediated (cross-sectionally) the association between maltreatment exposure and reward sensitivity. |
| Marusak et al. (2017) | 86 (43 MT, 43 NMT) |
12 (7–17) |
65% | 45% B 34% W 13% O 8% n/a |
PR and/or SR CAT | RRC | MT and NMT did not differ in symptoms severity. Cut-off clinical criteria n/a. | ↓ VTA - HPC ↑ S. Nigra - AMG6 |
|
| Mishra et al. (2020) | 45 MT | (10–18) | 33% | n/a | IR and SR CAT |
RRC | Participants did not meet clinical cut-off. | ↓ dACC - AI/FO | Post intervention, the internal attention training was linked to enhanced dACC – AI/FO rsFC. Post intervention, across all intervention arms, increased dACC – AI/FO rsFC was associated with improvement in academic performance as well as hyperactivity and inattention symptoms. |
| Rakesh et al. (2021)7 | 130 | 16 (15–18) |
51% | 91% W 9% A |
SR CAT |
NBS | At baseline 19% met clinical cut-off criteria for depression and 6% for anxiety. | At 16 years: ↑ Subcortex8 - Visual Network ↑ Subcortex8 - Default Mode Network ↑ Subcortex8 - Limbic Network ↑ Subcortex8 - Dorsal Attention Network ↑ Default Mode Network - Salience Network At 19 years: ↑ Default Mode Network – Visual Network ↑ Subcortex8 - Visual Network ↑ Default Mode Network - Dorsal Attention Network ↑ Subcortex8 – Dorsal Attention Network ↑ Default Mode Network - Salience Network ↑ Limbic Network – Visual Network |
Maltreatment exposure was associated with a longitudinal (from 16 to 19 years) increase in overall rsFC connectivity. This in turn was predictive of subsequent depression symptoms. |
| Rakesh et al. (2021)7 | 130 | 16 | 51% | 91% W 9% A |
SR CON |
RRC | At baseline 19% met clinical cut-off criteria for depression and 6% for anxiety. | At 16 years: ↓ within an extended Salience Network9 (for abuse only) At 19 years: ↑ within an extended Salience Network9 (for neglect only) |
Maltreatment exposure was associated with longitudinal increase in rsFC within an extended Salience Network9; this in turn predicted future symptoms of depression and substance misuse. Maltreatment exposure was associated with longitudinal increase in rsFC within an extended Default Mode Network10 only in males. |
| Saxbe et al. (2018) | 21 | 17 (15–19) |
43% | 38% H 33% W 10% B 10% A 10% 0 |
SR and PR11 CON |
RRC | Externalising and Internalising measures included as variables of interest. Cut-off clinical criteria n/a. | ↑ AMG - vACC/sgACC ↑ AMG - PCC |
Externalizing behaviour in mid-adolescence (mean age = 16) mediated the association between family aggression in early adolescence (mean age = 13) and AMG-vACC/sgACC rsFC in late adolescence (mean age = 17) |
| Silveira et al. (2020)12 | 392 | 17 (12–22) |
55% | 71% W 17% B 6% A 6% O |
SR CON |
RRC13 | At baseline 8% met clinical cut-off for depression and 1% for an anxiety disorder. | Mediators between maltreatment exposure and executive functioning: ↓ dACC - Occipital Lobe (CUN, LGG, CaS) ↓ dACC - Sensorimotor Network (PrCG, PoCG) ↓ dACC - STG ↓ INS (AI) - PFC (dlPFC and IFG/OFC) ↓ INS (AI) - Sensorimotor Network (PrCG, PoCG) ↓ IPS - Occipital Lobe (middle and superior occipital cortex) |
Patterns of reduced rsFC between salience network, motor network, PFC and occipital lobe mediated the relationship between childhood maltreatment exposure and poorer executive functions. These patterns of reduced rsFC longitudinally predicted high-risk drinking behaviour. |
| Silveira et al. (2021)12 | 475 | 17 (12–22) |
52.4% | 73% W 14% B 6% A 7% O |
SR CAT14 |
RRC | Most participants scored below clinical cut-off threshold15. | ↓ dACC - (AI + THA + aPFC) | Reduced dACC - (AI + THA + aPFC) rsFC longitudinally mediated the link between child neglect and externalizing symptoms. |
| Thomason et al. (2015) | 42 (21 MT, 21 NMT) |
13 (9–15) |
69% | 48% B 33% W 7% H 12% n/a |
SR and PR CAT |
SBC | MT group showed greater symptoms than NMT. Cut-off clinical criteria n/a. | ↓ AMG - dACC ↓ AMG - INS (AI) ↓ AMG - vACC/sgACC ↓ AMG - OFC |
n/a |
| Wesarg et al. (2021) | 774 | 19 (17–23) |
51% | n/a | SR CAT |
SBC and RRC | 15% met clinical cut-off for any psychopathology. | ↑ AMG – INS (PIns) | ↑ AMG – INS (PIns) rsFC was found only among TT carriers of rs1360780 (a single nucleotide polymorphism (SNP) within the FKBP5 gene). |
| Xu et al. (2020) | 100 (46 MT, 54 NMT) |
24 (n/a) |
59% | 100% A16 | SR CAT |
SBC | Participants above clinical cut-off were excluded. | ↓ HPC - EC | Reduced HPC - EC rsFC mediated (cross-sectionally) the association between maltreatment severity and poorer visual memory |
| Zhao et al. (2021) | 138 (65 MT, 73 NMT) |
21 (n/a) |
50% | 100% A16 | SR CAT |
ICA | Participants above clinical cut-off were excluded. | ↑ mPFC – (vACC + mPFC) ↑ INS – (AI + dACC17) ↓ IPL18 – (IPL + dlPFC) ↓ (AI + dACC17) – (PCC + PCu + AG) ↓ (PCC + PCu + AG) - (IPL18 + dlPFC) |
Among individuals exposed to early adversity, maltreatment severity was linked with increased mPFC – (vACC + mPFC) rsFC and decreased (PCC + PCu + AG) - (IPL18 + dlPFC) rsFC. |
| Zielinski et al. (2018) | 36 (17 MT, 19 NMT) |
14 (12–16) |
100% | 72% W 22% B 3% H 3% O |
SR CAT |
SBC | In the MT group: 41% had PTSD, 23–29% had an internalizing disorder, 11% drug or alcohol abuse. In the NMT participants did not meet clinical cut-off criteria. | ↓ AMG - vACC/sgACC ↓ vACC/pgACC - OFC/vmPFC ↓ vACC/pgACC - PCu ↓ dACC - PCu ↓ dACC - AG |
n/a |
Abbreviations Sample size: MT = maltreated sample; NMT = non-maltreated sample. Ethnicity: A = Asian; B = Black/African American; H = Hispanic/Latino; O = Other or Mixed; W = White/Caucasian/European. Maltreatment Assessment: CAT = categorical; CON = continuous; EXP = experimental; IR = institutional record; PI = previously institutionalised; PR = parent report; SR = self-report. Analytic Method: GBNA = graph-based network analysis; ICA = independent component analysis; graph-based network analysis; NBS = network-based statistics; ROI = region of interest; RRC = ROI-to-ROI connectivity; SBC = seed-based whole-brain connectivity. Concurrent psychopathology: EXC = excluded if met clinical threshold. BCT = all below clinical threshold. Connectivity Findings: AG = angular gyrus; AI = anterior insula; AMG = amygdala; aPFC = anterior/frontopolar prefrontal cortex; BA = Brodmann area; CaS = calcarine sulcus; CUN = cuneus; dlPFC = dorsolateral prefrontal cortex; dmPFC = dorsomedial prefrontal cortex; EC = Entorhinal Cortex; FEF = frontal eye fields FO = frontal operculum; GF = fusiform gyrus; HG = Heschl’s gyrus; HPC = hippocampus; INS = Insula; IPS = intraparietal sulcus; IPL = inferior parietal lobule; ITG = Inferior temporal gyrus; LGG = lingual gyrus; lPFC = lateral prefrontal cortex; MI = middle insula; OFC = orbitofrontal cortex; PCC = posterior cingulate cortex; PCu = precuneus; PHC = parahippocampal cortex; PIns = posterior insula; pgACC = pregenual anterior cingulate cortex; PoCG = postcentral gyrus; PrCG = precentral gyrus; sgACC = subgenual anterior cingulate cortex; SMG =supramarginal gyrus; STG = superior temporal gyrus; THA = thalamus; vACC = ventral anterior cingulate cortex; vmPFC = ventromedial prefrontal cortex; VTA = ventral tegmental area. Other: n/a = not available or not applicable.
1 Measured at baseline/first scan
2 main effect of maltreatment exposure (measured as a categorical or continuous variable)
3 MT and NMT participants were recruited from three different studies characterised by high risk/adversity exposure
4 MT participants show positive rsFC, while NMT participants show negative rsFC
5 the country of origin of participants was reported, but no ethnic information was provided
6 approached statistical significance (p = .056)
7Rakesh et al. (2021) and Rakesh et al. (2021) used the same sample in both studies
8 subcortex = thalamus, basal ganglia, amygdala, hippocampus
9 extended Salience Network = AI, dACC, dmPFC/SMA, lPFC, cerebellum
10 extended Default Mode Network = mPFC, AG, PCC, PCu, THA, HPC
11 maltreatment exposure was assessed only in the past year
12Silveira et al. (2020) and Silveira et al. (2021) used samples from the NCANDA cohort
13 multiple ROI (236) were selected for this RRC analysis, and only altered coupling between ROI that mediated the relationships between a history of maltreatment and poor executive functioning were reported
14 MT group score in the low-to-moderate range on CTQ
15 At baseline, 2 SD above the mean score for both internalising and externalising symptoms’ was below clinical cut-off
16 participants were Chinese (other ethnicity information was nor reported)
17 the Salience Network (i.e. AI + dACC) includes other adjacent regions, such as the mPFC and operculum as shown in Zhou et al. (2021) Fig. 1
18 the IPL in the original article this is referred to as inferior parietal gyrus.