Abstract
Bipolar disorder (BD) is a major psychiatric illness which is thought to have strong biological underpinnings. A biological basis for BD is exemplified by a strong heritability of the disorder (1), occurrence of mood periods of mania (BPM), and depression (BPD), which may or may not be precipitated by environmental factors and dramatic improvement with specific medication treatment such as lithium(2). Therefore, with the augment of brain imaging techniques to study brain metabolism and task-induced activation there is an expectation that a brain state or trait abnormalities specific to BD will be identified.
7.1 Introduction
Bipolar disorder (BD) is a major psychiatric illness which is thought to have strong biological underpinnings. A biological basis for BD is exemplified by a strong heritability of the disorder (1), occurrence of mood periods of mania (BPM), and depression (BPD), which may or may not be precipitated by environmental factors and dramatic improvement with specific medication treatment such as lithium(2). Therefore, with the augment of brain imaging techniques to study brain metabolism and task-induced activation there is an expectation that a brain state or trait abnormalities specific to BD will be identified. Indeed, regional brain abnormalities in the orbitofrontal cortex, anterior cingulate cortex, striatum, thalamus, and amygdala have been identified in BD; however, no particular abnormality has been consistently reported. In light of this state of affairs, it has been proposed that the abnormality in BD may lie in the functional connectivity (FC) between brain regions rather than in a particular region (3).
7.2 The Functional Connectome
The concept of functional connectome has been developed recently to denote the functional connectivity between brain regions. Functional connectivity is a correlation of neuronal activity between brain regions during task-induced activation or at rest. The former is, however, more accurately labeled as task-related co-activation or coupling of activation as only that activity which changes in response to an active task versus a control task is measured. Functional connectivity is a correlation between two brain regions and as such does not provide any direct knowledge of the effect of one region or the other. Functional connectivity between two regions can occur because of one region influencing the activity of the other or vice versa or both or even by another factor or region simultaneously affecting both the regions in question (4). Effective connectivity, on the other hand, implies the effect of the activity of one region over that of another region (4). Techniques such as dynamic causal modeling (5) have been developed to measure effective connectivity; however, they are more cumbersome to measure and have not been as much investigated as functional connectivity. The most commonly used methods to study the functional connectome are described later.
7.2.1 Methods to Study the Functional Connectome in Brain Imaging Studies
Task-Based Functional Connectivity:
A number of studies have investigated the coupling of activation while subjects perform an activation task. As noted earlier, this is not strictly a measure of functional connectivity as activation depends on the baseline level of activity of a region that can influence whether a higher or lower BOLD change is observed. Therefore, mere correlation of time series during an activation task should not be done. A more accurate measure of co-activation is pathophysiological interaction (PPI) in which activation in two regions is correlated in the context of a task (6). Several studies have used PPI to investigate functional connectivity in BD.
Brain task-related functional connectome findings in bipolar disorders are depicted in Table 7.1.
Task | Study | Subjects | Medication | Area | BPE/BP vs. HC | BPM vs. BPD vs. BPE vs. HC | BPD vs. MDD |
---|---|---|---|---|---|---|---|
|
|
| dlPFC-amygdala | ↑ BPE compared with HC | ↑ BP II compared with BPE I and HC | ||
Facial emotion identification/matching tasks Emotional regulation task | 14 BP and 14 HC 30 BPE I and 26 HC |
| vPFC-amygdala | ↓ BP compared with HC (+) | |||
Facial emotion identification/matching tasks | Wang (81) | 33 BP and 31 HC | Combinations | pACC-amygdala | ↓ BP compared with HC | ||
CPT-END task | Cerullo (11) | 15 BP I and 15 HC | Combinations | IFC-amygdala | ↑BPM | ||
CPT-END task | Cerullo (11) | 15 BP I and 15 HC | Combinations | insula-right amygdala | ↑ Transition from BPM to BPD | ||
During sad experiment, facial emotion identification/matching tasks | Versace (12) | 31 BP and 24 HC | Combinations | OFC-amygdala | ↑ BP compared with HC | ||
During happy experiment, facial emotion identification/matching tasks | Versace (12) | 31 BP and 24 HC | Combinations | OFC-amygdala | ↓ BP compared with HC | ||
Verbal working memory task | Stegmayer (13) | 18 BPE and 18 HC | Combinations | right hemispheric-amygdala | ↓ BPE compared with HC | ||
2-Back working memory task | Goikolea (14) | 31 FEM and 31 HC | Combinations | vmPFC-SFG | ↑ BPM compared with HC | ||
Motor activation paradigm | Marchand (15) | 19 BPE II and 18 HC | Combinations | within SFG | ↑ BPE compared with HC | ||
During reward receipt reward processing task | Dutra (16) | 24 BP I and 25 HC | Combinations | OFC-VS | ↑ BP compared with HC | ||
|
|
| PFC-VS | ↓ BP compared with HC (+) | |||
Card-guessing paradigm | Redlich (18) | 33 BP, 33 MDD, and 34 HC | Combinations | VTA-VS | ↑ MDD compared with HC (Classification between MDD and BP: 66.6% accuracy) | ||
Distraction and reappraisal emotion regulation task | Lois (19) | 21 BP, 21 MDD, and 23 HC | Combinations | within DMN, between DMN and CCN | ↑ Both BPD and MDD |
Symbols: +, more than one report from different investigators; negative or contrary finding reported; ↑, increased functional connectivity; ↓ decreased functional connectivity
Disorders: BP, bipolar; BPE, euthymic bipolar; BPM, manic bipolar; BPD, depressive bipolar; MDD, major depressive disorder; HC, healthy control; BP I, bipolar type I; BP II, bipolar type II; BPE I, euthymic bipolar type I; BPE II, euthymic bipolar type II; FEM, first episode mania
Brain regions: PFC, prefrontal cortex; dlPFC, dorsolateral PFC; vPFC, ventral PFC; vlPFC, ventrolateral PFC; vmPFC, ventromedial PFC; aPFC, anterior PFC; mPFC, medial PFC; IFC, inferior frontal cortex; OFC, orbitofrontal cortex; mOFC, medial OFC; ACC, anterior cingulate cortex; pACC, perigenual ACC; sgACC, subgenual ACC; PCC, posterior cingulate cortex; IFG, inferior prefrontal gyrus; ITG, inferior temporal gyrus; SFG, superior frontal gyrus; VTA, ventral tegmental area; SMA, supplementary motor area; VS, ventral striatum
Brain networks: DMN, default mode network; CCN, cognitive control network; SMN, sensorimotor network; CEN, central executive network; SN, salience network
Analysis: ICA, independent component analysis
Task: CPT-END, continuous performance task with emotional and neutral distractors
Resting-State Low-Frequency BOLD Fluctuations Correlation (Connectivity):
As discussed earlier, computing functional connectivity from task-induced co-activation has some limitations. Resting-state low-frequency BOLD fluctuations (LFBF) correlation has provided a much more powerful and relatively easy to measure method to deduce functional connectivity between brain regions. Resting-state functional connectivity (RSFC) has been empirically shown to correlate between brain regions known to have functional or anatomical connections (20–22). Three commonly used methods to quantify results from the resting-state LFBF correlation analysis are to study the functional connectome in health and disease – seed-based analysis, independent component analysis, and graph-theory analysis.
Brain resting-state functional connectome findings in bipolar disorders are depicted in Table 7.2.
Study | Subjects | Medication | Area | BPE/BP vs. HC | BPM vs. BPD vs. BPE vs. HC | BPD vs. MDD |
---|---|---|---|---|---|---|
|
| ACC-amygdala | ↑ BPE compared with HC |
| ||
|
| dlPFC-amygdala | ↓ BP compared with HC | ↑ BPM compared with BPE | ||
|
| (v or m)PFC-amygdala |
| ↓ Both BPM and BPD | ||
|
| OFC-amygdala |
| |||
Singh (31) | 20 BP I, 23 HC | Combinations | preceuneus-amygdala | ↓ BP I compared to HC | ||
Li (32) | 21 BPE and 28 HC | Lamotrigine for BPE | SMA-amygdala | ↓ BPE compared with HC | ||
Chen (33) | 43 BP II, 36 MDD and 47 HC | Unmedicated | ACC-OFC | ↓ Both MDD and BPD compared with HC | ||
Magioncalda (34) Gong (35) |
| Combinations Unmedicated | pACC-ITG | ↓ BP compared with HC (+) | ||
|
| ACC-PCC | ↓ BP compared with HC | ↓ BPM compared with Both BPD and HC | ||
He (38) | 32 BP, 33 MDD, and 43 HC | Combinations | dlPFC-cerebellar | ↓ BPD compared with Both MDD and HC (Classification between BPD and MDD : 69% accuracy) | ||
Favre (26) | 20 BPE and 20 HC | Combinations | mPFC-dlPFC | ↑ BP compared with HC | ||
Gong (35) | 96 BP II and 100 HC | Unmedicated | mPFC-PCC | ↓ BP compared with HC | ||
Minuzzi (39) | 32 right-handed BPE women and 36 HC | Combinations | OFC-IFG | ↑ BPE compared with HC | ||
Wang (40) | 36 BP II, 32 MDD, and 40 HC | Combinations | inter-hemispheric | ↓ BP II compared with HC | ||
Yasuno (41) | 16 BP and 22 MDD | Combinations | inter-hemispheric | ↓ BPD compared with HC | ||
Reinke (42) | 21 BPE and 20 HC | Combinations | IFG-Insula | ↑ BP compared with HC | ||
Li (32) | 18 BPM, 10 BPD and 28 HC, 26 BPM, 21 BPE | Combinations | IFG-lingual gyrus | ↓ Both BPM and BPD | ||
|
| insula-inferior parietal lobe | ↑ BPD compared with MDD (−) | |||
Minuzzi (39) | 32 right-handed BPE women and 36 HC | Combinations | insular-somatosensory cortex | ↑ BPE compared with HC | ||
Marchand (45) | 14 BPD and 26 MDD | Combinations | PCC-inferior parietal lobule, precentral gyrus and insula | ↑ BPD compared with MDD | ||
Yin (46) | 21 BP, 40 MDD and 70 HC | Combinations | SFG-insula | ↓ BPD compared with HC | ||
Pang (43) | 30 BP, 30 MDD, and 30 HC | Combinations | dlPFC-Insula | ↑ BPD compared with MDD | ||
Liu (47) | 17 BP and 17 MDD | Combinations | IFG-hippocampal | ↑ BPD compared with MDD, ↑ Both BPD and MDD compared with HC | ||
Fateh (48) | 30 BPD, 29 MDD, 30 HC | Combinations | lingual gyrus-hippocampal | ↑ BPD compared with MDD | ||
Chen (33) | 43 BP II, 36 MDD and 47 HC | Unmedicated | SFG-hippocampal | ↓ MDD compared with HC | ||
Oertel-Knochel (49) | 21 BP and 21 HC | Combinations | IFG-hippocampal | ↑BP compared with HC | ||
Dandash (50) | 61 FEM and 30 HC | Combinations | in the dorsal and caudal cortico-striatal systems | ↓ BPM compared with HC | ||
Anand (3) | 6 BDM, 5 BDD, 15 MDD and 15 HC | Unmedicated | pACC-striatum | ↓ Both BPM and BPD | ||
Anand (3) | 6 BDM, 5 BDD, 15 MDD and 15 HC | Unmedicated | pACC-thalamus | ↓ Both BPM and BPD | ||
He (51) | 25 BPD, 25 MDD, and 34 HC | Combinations | dlPFC-striatum | ↑ BPD compared with MDD | ||
Altinay (52) | 30 BDD, 30 BDM, and 30 HC | Unmedicated | left dorsal caudate and midbrain regions | ↑ BPM | ||
Altinay (52) | 30 BDD, 30 BDM, and 30 HC | Unmedicated | caudate-midbrain region | ↑ BPM | ||
|
| hippocampus-amygdala | ↑ BP I compared with HC | ↓ MDD compared with Both BPD and HC | ||
Teng (54) | 15 BP I, 16 HC | Combinations | thalamic–hippocampus | ↑ BP I | ||
Dandash (50) | 61 FEM and 30 HC | Combinations | VS-thalamus | ↑ BPM | ||
| Combinations Combinations | language areas | ↑ BP compared with HC | ↓ BPD compared to BPE | ||
|
| the cerebellar crus and lobules with areas of the frontal cortex | ↓ BP II compared with HC (++) | |||
Shi (60) | 66 BPD and 40 HC | Combinations | VTA-VS | ↓ BPD compared with MDD (Classification between BPD and MDD: 70% accuracy) | ||
Han (61) | 40 BPD, 54 MDD and 44 HC | Combinations | raphe nucleus with subcortical regions | ↓ BPD |
Symbols: +, more than one report from different investigators; −, negative or contrary finding reported; ↑, increased functional connectivity; ↓ decreased functional connectivity
Disorders: BP, bipolar; BPE, euthymic bipolar; BPM, manic bipolar; BPD, depressive bipolar; MDD, major depressive disorder; HC, healthy control; BP I, bipolar type I; BP II, bipolar type II; BPE I, euthymic bipolar type I; BPE II, euthymic bipolar type II; FEM, first-episode mania
Brain regions: PFC, prefrontal cortex; dlPFC, dorsolateral PFC; vPFC, ventral PFC; vlPFC, ventrolateral PFC; vmPFC, ventromedial PFC; aPFC, anterior PFC; mPFC, medial PFC; IFC, inferior frontal cortex; OFC, orbitofrontal cortex; mOFC, medial OFC; ACC, anterior cingulate cortex; pACC, perigenual ACC; sgACC, subgenual ACC; PCC, posterior cingulate cortex; IFG, inferior prefrontal gyrus; ITG, inferior temporal gyrus; SFG, superior frontal gyrus; VTA, ventral tegmental area; SMA, supplementary motor area; VS, ventral striatum
Brain networks: DMN, default mode network; CCN, cognitive control network; SMN, sensorimotor network; CEN, central executive network; SN, salience network
Analysis: ICA, independent component analysis
Task: CPT-END, continuous performance task with emotional and neutral distractors
Seed-Based Functional Connectivity Analysis:
For this method, a reference area of interest is first identified and then the mean resting-state BOLD fluctuations of this region are correlated with the mean of BOLD fluctuations in one or more target regions of interest (ROIs) or all of the voxels of the whole brain. A majority of studies that have hypothesized an a priori reference ROI have been conducted using the ROI approach.
Independent Component Analysis (ICA):
Independent component analysis of the resting-state BOLD signal has revealed several components comprising correlated brain regions which have been named according to their purported neuropsychiatric function – default mode, salience, executive function, and others (62, 63). The ACC and PCC connectivity is part of a default mode network (DMN) that shows high connectivity during rest, and these areas get deactivated when any task is conducted (64). The default mode circuit has been related to consciousness and vigilance to external and internal milieus while no task is being conducted (65). The salience network (SN) comprising the insula and other cortical areas is thought to be involved in the assessment of the internal mental and emotional state, while executive motor network (EMN) comprises correlated motor areas.
Brain functional connectome findings in bipolar disorders analyzed by independent component analysis are depicted in Table 7.3.
Table 7.3 Brain functional connectome findings in bipolar disorders analyzed by independent component analysis
Study | Subjects | Medication | Area | BPE/ BP vs. HC | BPM vs. BPD vs. BPE vs. HC | BPD vs. MDD |
---|---|---|---|---|---|---|
Ishida (66) | 22 BP and 24 HC | Combinations | In two clusters in the SMN (right and left primary somatosensory areas) | ↓ BP compared with HC | ||
Syan (67) | 32 BPE women and 36 HC | Combinations | PCC-angular gyrus | ↑ BPE | ||
Lois (68) | 30 BDE I and 35 HC | Combinations | Between the meso/paralimbic and the right frontoparietal network | ↑ BPE compared with HC | ||
Lois (68) | 30 BDE I and 35 HC | Combinations | Across the bilateral insula and putamen and across a temporo-insular network | ↑ BP II | ||
Martino (37) | 20 BPD, 20 BPM, 20 BPE, and 40 HC | Combinations | Within the DMN and SMN | ↑ BPD compared with BPM | ||
Ford (69) | 15 BPD and 15 MDD | Combinations | In ICA components | ↓ Both BPD and MDD | ||
Wang (59) | 38 BPD, 35 MDD, and 47 HC | Unmedicated | Intra-network FC within the DMN | ↓ Both the BPD and MDD | ||
Wang (59) | 38 BPD, 35 MDD, and 47 HC | Unmedicated | Inter-network FC between the CEN and SN | ↑ BPD compared with either the MDD or HC | ||
He (70) | 13 BP, 40 MDD, and 33 HC | Unmedicated | Within sensory, motor and cognitive networks | ↓ BP compared with MDD (Classification between BP and MDD : 99% accuracy) | ||
Goya-Maldonado (71) | 20 BPD, 20 MDD, and 20 HC | Combinations | In the frontoparietal network | ↑ BPD | ||
Goya-Maldonado (71) | 20 BPD, 20 MDD, and 20 HC | Combinations | in the DMN | ↑ MDD |
Symbols: +, more than one report from different investigators; −, negative or contrary finding reported; ↑, increased functional connectivity; ↓ decreased functional connectivity
Disorders: BP, bipolar; BPE, euthymic bipolar; BPM, manic bipolar; BPD, depressive bipolar; MDD, major depressive disorder; HC, healthy control; BP I, bipolar type I; BP II, bipolar type II; BPE I, euthymic bipolar type I; BPE II, euthymic bipolar type II; FEM, first-episode mania
Brain regions: PFC, preFrontal cortex; dlPFC, dorsolateral PFC; vPFC, ventral PFC; vlPFC, ventrolateral PFC; vmPFC, ventromedial PFC; aPFC, anterior PFC; mPFC, medial PFC; IFC, inferior frontal cortex; OFC, orbitofrontal cortex; mOFC, medial OFC; ACC, anterior cingulate cortex; pACC, perigenual ACC; sgACC, subgenual ACC; PCC, posterior cingulate cortex; IFG, inferior prefrontal gyrus; ITG, inferior temporal gyrus; SFG, superior frontal gyrus; VTA, ventral tegmental area; SMA, supplementary motor area; VS, ventral striatum
Brain networks: DMN, default mode network; CCN, cognitive control network; SMN, sensorimotor network; CEN, central executive network; SN, salience network
Analysis: ICA, independent component analysis
Task: CPT-END, continuous performance task with emotional and neutral distractors
Graph-Theory Analysis:
Graph-theory metrics provide measures of network-wide properties to provide insights into network function rather than the strength of connectivity between seed regions (72). Nodes and edges (connections) between nodes are measured in terms of network organization patterns related to network Resilience (e.g., assortativity), Segregation (e.g., clustering coefficient, transitivity), Integration (e.g., diffusion efficiency), and Centrality (e.g., pageRank centrality, subgraph Centrality).
Brain functional connectome findings in bipolar disorders analyzed by graph-theory method are depicted in Table 7.4.
Study | Subjects | Medication | Area | BPE/ BP vs. HC | BPM vs. BPD vs. BPE vs. HC | BPD vs. MDD |
---|---|---|---|---|---|---|
Doucet (73) | 78 BP, 64 unaffected siblings, and 41 HC | 78 BP, 64 unaffected siblings, and 41 HC | Global cohesiveness and their un affected siblings | ↓ BP and their unaffected siblings compared with HC | ||
Wang (74) | 37 BP II and 37 HC | Unmedicated | DMN | ↓ BP compared with HC | ||
Wang (74) | 37 BP II and 37 HC | Unmedicated | Limbic regions | ↑ BP II compared with HC | ||
Spielberg (75) | 30 BPM, 30 BPD and 30 HC | Unmedicated | Amygdala centrality | ↑BPM | ||
Spielberg (75) | 30 BPM, 30 BPD and 30 HC | Unmedicated | OFC centrality | ↓BPD | ||
He (76) | 13 BP, 40 MDD and 33 HC | Unmedicated | ICA components | ↑ BPD compared with MDD | ||
Wang (77) | 31 BP II depressed, 32 MDD, and 43 HC | Unmedicated | In the bilateral precuneus | ↓ Both BPD and MDD short range FCS | ||
Wang (77) | 31 BP II depressed, 32 MDD, and 43 HC | Unmedicated | In the bilateral cerebellum | ↑ BPD compared with MDD long-range FCS and short range FCS | ||
Wang (77) | 31 BP II depressed, 32 MDD, and 43 HC | Unmedicated | In the DMN, limbic network and cerebellum | ↓ Both the MDD and BP II nodal characteristics (nodal strength and nodal efficiency) |
Symbols: +, more than one report from different investigators; negative or contrary finding reported; ↑, increased functional connectivity; ↓ decreased functional connectivity
Disorders: BP, bipolar; BPE, euthymic bipolar; BPM, manic bipolar; BPD, depressive bipolar; MDD, major depressive disorder; HC, healthy control; BP I, bipolar type I; BP II, bipolar type II; BPE I, euthymic bipolar type I; BPE II, euthymic bipolar type II; FEM, first-episode mania
Brain regions: PFC, prefrontal cortex; dlPFC, dorsolateral PFC; vPFC, ventral PFC; vlPFC, ventrolateral PFC; vmPFC, ventromedial PFC; aPFC, anterior PFC; mPFC, medial PFC; IFC, inferior frontal cortex; OFC, orbitofrontal cortex; mOFC, medial OFC; ACC, anterior cingulate cortex; pACC, perigenual ACC; sgACC, subgenual ACC; PCC, posterior cingulate cortex; IFG, inferior prefrontal gyrus; ITG, inferior temporal gyrus; SFG, superior frontal gyrus; VTA, ventral tegmental area; SMA, supplementary motor area; VS, ventral striatum
Brain networks; DMN, default mode network; CCN, cognitive control network; SMN, sensorimotor network; CEN, central executive network; SN, salience network
Analysis: ICA, independent component analysis
Task: CPT-END, continuous performance task with emotional and neutral distractors
7.2.2 The Functional Connectome in Bipolar Disorder
7.2.2.1 Study Designs for Investigation of Functional Connectome in Bipolar Disorder
Since we published the first report of abnormalities of resting-state functional connectivity in major depression (MDD) and BD (3, 22), the number of reports of FC abnormalities in BD has exponentially increased. Various studies have used several different experimental strategies to investigate the functional connectome pathophysiology in BD. Comparison with healthy controls (HCs) and BD subjects is one straightforward strategy. Studies of euthymic subjects versus HCs provide evidence for trait-related abnormalities in BD. However, it is challenging to study truly euthymic bipolar subjects particularly in the absence of confounds such as medication load effects. Another strategy to study trait-related abnormalities is to study affected and unaffected relatives of BD subjects.
Studies investigating mania or depression provide information regarding state-related connectome abnormalities, though technically BPM and BPD have a combination of state- and trait-related abnormalities. Abnormalities common to both BPD and BPM could be thought of as trait-related abnormalities, or they could be common state-related abnormalities that can give rise to both mania and depression, for example, a general emotional dysregulation. A powerful strategy is to compare the different states rather than comparison with healthy subjects as other confounds related to bipolar illness can be controlled. An even more attractive approach is to study the various states within the same subject, though in this design, it is difficult to control for changes in confounding factors such as environmental, biological, and therapeutic factors as a subject transitions from one state to the other. The comparison with unipolar major depression (MDD) is not only a powerful strategy to isolate abnormalities related to bipolarity but also a highly clinical relevant distinction for which a biomarker is critically needed.
For the purpose of this review, we examined all reports in which trait and state-related abnormalities of the FC have been investigated. Pediatric BD and comparison with schizophrenia and other non-mood disorders studies are not included as that is beyond the scope of this review. Several findings related to different brain regions have been reported; therefore we have organized them into the major categories of hypothesized FC abnormalities in BD – cortico-limbic, cortico-cortical, and subcortical for reference ROI methodology, ICA component abnormalities, and graph-theory property abnormalities. We discuss each of these abnormalities reported in the context of the comparison group experimental paradigm used.
7.2.2.2 ROI Based Analysis
7.2.2.2a Cortico-Limbic Connectivity
The study of cortico-limbic connectivity implies that investigators studied the relationship between cortical mood-regulating areas, for example, anterior cingulate cortex (ACC), orbitofrontal cortex (OFC), and subcortical or limbic mood-generating areas such as the amygdala (AMYG) and ventral striatum (VS). This hypothesis is derived from the Jacksonian view of neural architecture, which models the brain in terms of hierarchical structures with the higher structures influencing the activity of the lower structures (79). The influence of mood-regulating regions on the mood-generating regions also harkens back to the Freudian model of the psyche in which the superego and ego control the expression of the instinctual drives of the id. It should be noted though that despite being a neurologist, Freud did not propose a neuroanatomical model for his model of the organization of the psyche (80). As BD involves impairments in inhibition of emotional responses and impulsive behavior, it is thought that cortico-limbic emotion regulatory mechanisms have become impaired in some way leading to unregulated mood symptomatology. As noted earlier, functional connectivity does not give any information regarding the effect of one region over the other but coherence of activity between brain regions has been thought to provide some information regarding whether two brain areas are working simultaneously or not. The findings of altered cortico-limbic functional connectivity in BD are summarized later.
Task-Related Cortico-Limbic Psychophysiological Interaction (PPI) Studies:
Using facial emotion identification/matching tasks, several studies have reported decreased connectivity between amygdala and areas of the frontal cortex such as the ventromedial prefrontal cortex (vmPFC) (9) and posterior anterior cingulate cortex (pACC) (81). On a verbal working memory task, euthymic BD subjects (BPE) exhibited decreased PPI right-sided cortico-amygdalar connectivity compared to HC (13). Other studies have reported increased cortico-limbic connectivity: increased PPI dlPFC-amygdalar connectivity in BPE versus HC using an emotional Stroop task (8) and during an emotion regulation task, PPI cortico-limbic connectivity has been reported to be less decreased in BPE I subjects compared to HC (7, 10). Still, other studies have reported both an increase and a decrease depending on the emotional task used: using a face processing task, increased pACC-amygdala connectivity during processing of sad faces versus decreased OFC-amygdala connectivity during processing of happy faces was reported in remitted/depressed BD subjects compared to HCs (12). During a reward-processing task, increased FC between VS and mPFC during reward receipt and decreased FC during reward omission in BD I subjects compared to controls(16), while in a reward anticipation task decreased FC between VS and anterior prefrontal cortex (aPFC) was seen (17).
In a study that investigated state-related differences, subjects who transitioned from BPM to BPD after treatment, BPM at baseline had increased frontal gyri (IFC)-amygdala correlation of activations during continuous performance task (CPT-END) while when they transition to a BPD state, this connectivity was decreased but right amygdala-insula connectivity increased(11). In another study in which BPE I and BPE II were directly compared in terms of PPI FC during an emotion regulation task, BPE I subjects exhibited a decreased inverse correlation between cortico-amygdalar connectivity while BP-II subjects exhibited increased inverse correlation suggesting a difference between the two subtypes (7).
In studies that investigated differences between BPD and MDD, the following cortico-limbic FC abnormalities have been reported: Redlich and colleagues using PPI during a card-guessing reward-processing task reported increased FC between the VS and ventral tegmental area (VTA) in the MDD group compared to HC, but no differences with BPD group were found compared to HC (18). In a study that investigated effective connectivity differences in amygdala-OFC connectivity in BD and MDD subjects, Almeida and colleagues reported top-down left-sided amygdala-OMPFC abnormality in MDD and right-sided bottom-up abnormality in BD (82).
Resting-State Reference ROI-Based Cortico-Limbic Connectivity Studies:
Anand and colleagues first reported decreased resting-state LFBFs correlation with perigenual ACC-limbic connectivity in medication-free subjects compared to HCs in both BPD and BPM groups compared to HCs (3). Since the first report, several studies have studied cortico-limbic, particularly cortico-amygdalar connectivity in BD and in general have reported decreased amygdala connectivity with dorsal PFC areas but increased amygdala and other limbic areas with the ventral PFC in BD. Chepenik and colleagues reported increased vPFC correlations with amygdala (28). Increased amygdala-mPFC but decreased amygdala-dlPFC connectivity was also reported in remitted BD patients with and without psychosis (25). Increased frontal-hippocampal and vlPFC-VS FC has also been reported in BD subjects compared to HCs (49). Studies that have specifically looked at differences between BPE and HC to identify trait-related abnormalities have reported hyperconnectivity between right amygdala and right vlPFC (29), decreased amygdala connectivity with supplementary more area (32), greater connectivity between mPFC and right amygdala compared to HS, which was also correlated with the duration of the disease (26), and increased amygdala connectivity to the subgenual ACC (23).
A number of cortico-limbic RSFC studies have looked at state-related differences between bipolar mood states. In general, both BPD and BPM have been found to share cortico-limbic connectivity abnormalities, but some differences were also found. Decreased pregenual ACC connectivity with the striatum and thalamus (3), decreased amygdala connectivity with inferior frontal orbital gyrus and lingual gyrus (30), decreased FC between the amygdala and left middle frontal cortex (27), and widespread common cortico-striatal connectivity abnormalities (52) have been reported in both states. Decreased FC between right OFC and amygdala in BPM compared to BPD and in a study comparing BPM with BPE subjects, decreased connectivity between amygdala and ACC in BPM has been reported (24). Conversely, increased connectivity between the amygdala and dorsal frontal cortical structures involved in emotion regulation has also been observed (24). Cortico-striatal connectivity has been reported to be different in BPM and BPD subjects. In a relatively large study comparing medication-free subjects, BPD showed increased connectivity of the dorsal caudal putamen with somatosensory areas such as the insula and temporal gyrus while BPM showed unique increased connectivity between left dorsal caudate and midbrain regions as well as increased connectivity between VS and thalamus (52). Another study reported similar findings in first episode manic patients regarding reduced connectivity in the dorsal and caudal cortico-striatal systems and increased connectivity in a circuit linking the VS with the medial orbitofrontal cortex, cerebellum, and thalamus when compared to HCs (50).
Studies with an aim to distinguish between BPD and MDD have reported significant differences in hippocampal and striatal FC. In an FDG PET study, Benson and colleagues (83) reported an increased correlation between hippocampus and prefrontal areas in MDD versus BPD. In resting-state LFBF studies also, hippocampal connectivity abnormalities have been found in both BPD and MDD with some differences. In BPD patients, increased FC of the bilateral anterior/posterior hippocampus with lingual gyrus and inferior frontal gyrus (IFG) relative to MDD patients was observed while in comparison to HCs, both groups had an increased FC between the right anterior hippocampus and lingual gyrus and a decreased FC between the right posterior hippocampus and right IFG (47). Increased FC between IFG and lingual gyrus with the hippocampus in BPD compared to MDD (48) was also observed. Increased FC of the striatum in BPD versus MDD has also been reported in a few studies. Increased positive metabolic correlations between prefrontal and ventral striatal areas(83), increased FC between VS and ACC (84), and increased dlPFC connectivity with the striatum in BPD compared to MDD in a PET cerebral blood flow study (51) have been reported.

Stay updated, free articles. Join our Telegram channel

Full access? Get Clinical Tree


