Inflammatory Biomarkers, Cognitive Functioning, and Brain Imaging Abnormalities in Bipolar Disorder: A Systematic Review

Abstract

Objective

Recent studies have pointed to neuroinflammation and neurotrophic factors as crucial mediators in the pathophysiology origins of mood disorders. The aim of this review is to assess the potential association between cognitive impairment, brain imaging abnormalities, and inflammatory biomarkers in patients affected by bipolar disorder (BD).

Method

Following PRISMA (Preferred Reporting Items for Systematic reviews and Meta-analyses) guidelines, we systematically searched PubMed, Google Scholar, Scopus, and Web of Science databases, with no year restriction, up until August 2023, for human studies that examined the relationship between inflammatory markers and cognitive impairment in BD patients. Studies based on neuroimaging, such as MRI, DTI, and fMRI, were also included, along with those examining the moderating role of specific inflammatory markers in the alteration of the brain.

Results

59 human clinical studies satisfied the criteria for consideration. Most of the studies reviewed concur that inflammatory state, measured by peripheral blood levels of CRP and cytokines, constitutes an important contributor to cognitive impairment observed in patients with BD. Robust evidence indicates an association between cognitive impairment and CRP, IL-1RA, IL-6, and TNF-α with its receptors, whereas there is no convincing evidence for the involvement of other neuroinflammatory biomarkers. Neuroimaging studies suggest that brain structural/functional abnormalities seen in BD could also be linked to a neuroinflammatory condition.

Conclusions

Current data provide evidence of a link between cognitive impairments observed in BD patients and mechanisms of neuroinflammation. Emerging evidence indicates that systemic inflammation might also play an important role in the deterioration of brain structures critical to cognitive functions in patients with BD. The convergence of findings across these studies strengthens our understanding of the complex neurobiological underpinnings of these disorders. Identification of BD specific inflammatory markers may be of assistance for future early therapeutic interventions.

Introduction

Bipolar disorders (BD) are severe mental health disorders that cause dramatic changes in a person's mood, energy, and ability to function. Over recent decades cognitive impairment has been increasingly recognized as a core feature of BD because it persists during periods of clinical remission (euthymia) and predicts poor psychosocial functioning and lower quality of life in patients with BD (Barch, 2009; Cotrena et al., 2016; Vrabie et al., 2015). Although a pathognomonic profile of neurocognitive dysfunction has not been identified, numerous studies have demonstrated the presence of specific neuropsychological impairments including deficit in executive functioning, working memory, attentional processes, and verbal learning and memory (Bortolato et al., 2015; Cardenas et al., 2016; Vöhringer et al., 2013). The pathophysiology of cognitive impairment in BD remains largely unknown. Evidence strongly suggests that inflammation and immune dysfunction play a significant role in the aetiology of cognitive dysfunction (Stertz et al., 2013). Indeed, several studies have shown elevated levels of pro-inflammatory cytokines in BD, suggestive of chronic low-grade inflammation. Many studies have found that inflammation is likely a pertinent pathogenic factor for only a subset of BD. This subset of BD may represent an “inflammatory BD” that is pathophysiologically and phenomenologically different from other forms of bipolar disorder. Understanding the relationship between inflammation and cognitive impairment in BD may help to establish subgroups of patients associated with inflammatory activation and those at greater risk of cognitive decline (Goldsmith et al., 2016; Modabbernia et al., 2013; Munkholm et al., 2013; Pape et al., 2019). Serum levels of pro-inflammatory molecules interleukin-4 (IL4), tumour necrosis factor alpha (TNF-α), interleukin-6 (IL-6), interleukin-1beta (IL-1β), soluble IL-2 receptor, soluble receptor of TNF-α type 1 (sTNFR1) and C-reactive protein (CRP) are elevated in BD patients compared to healthy controls (Barbosa et al., 2014). Findings also suggest that cytokine levels vary depending on mood state (i.e., differing profiles during periods of mania and depression), with altered immune responses persisting even during euthymic periods (Brietzke et al., 2009; Brunoni et al., 2020; Modabbernia et al., 2013). Evidence to date implicates those inflammatory markers in cognitive dysfunction in BD, although this pertains to a limited set of biomarkers (Rosenblat et al., 2018). Several plausible mechanisms have been proposed to explain the relationship between inflammation and cognitive dysfunction in BD. Cytokines can directly alter monoamine levels, cause overactivation of microglia cells and lead to increased oxidative stress and hypothalamic–pituitary–adrenal (HPA) axis dysregulation in the brain. However, it is still unclear whether these mechanisms are associated with the risk of developing cognitive impairment in patients with BD. A clearer understanding of the association between immune dysfunction and cognitive impairments in BD may be of great interest with the potential for significant therapeutic implications (Green, 2006; Tse et al., 2014). Furthermore, patients with BD show a range of differences in functional and structural brain measures compared with healthy controls (Emsell & McDonald, 2009; Gong et al., 2021). The most consistent findings include regional grey and white matter structural changes and functional abnormalities in specific brain regions participating in mood and cognitive regulation, such as the prefrontal and limbic networks (Strakowski et al., 2005). The cause of these differences remains uncertain.

There is substantial evidence indicating that inflammation plays a role in structural and functional brain changes. This is because cytokines are capable of crossing the blood-brain barrier (BBB), leading to a cascade of effects such as cell death and reduced neurogenesis. These processes can consequently alter the brain's structure and functionality (Rosenblat et al., 2018). However, little is known about the mechanisms by which inflammatory cytokines and immune dysfunction affect brain structure (Bauer et al., 2014). Understanding the association between immune dysfunction and functional-structural brain abnormalities seen in BD would be a significant advance and allow a deeper understanding of causal pathways towards cognitive impairment in BD. While many systematic reviews of studies examining the association between inflammatory markers and cognition in BD have been published (e.g. Bauer et al., 2014; Misiak et al., 2018; Rosenblat et al., 2015), several relevant studies have been published subsequently. Additionally, these systematic reviews did not examine research investigating the link between inflammatory molecules and brain abnormalities, both structural and functional, as identified through neuroimaging techniques. Consequently, the aim of this systematic review is to consolidate existing research that delineates the connection between inflammation (evidenced by alterations in inflammatory markers) and immune dysfunction, and their association with cognitive deficits and neuroimaging anomalies in bipolar disorder (BD).

Methods

For this systematic review, PubMed, Google Scholar, Scopus, and Web of Science databases were searched, with no year restriction, for research articles addressing the relationship between BD, inflammation, cognitive impairment, and neuroimaging abnormalities. Search strings included a range of keyword combinations derived from the following terms: 'bipolar disorder', 'cytokines', 'interleukins', 'C-reactive protein', 'tumour necrosis factor', 'inflammation', 'inflammatory markers', 'immunological alterations', 'cognition', 'cognitive impairment', 'neuropsychological deficits', 'executive functions', 'memory', 'working memory', 'attention', 'imaging', 'magnetic resonance imaging (MRI)', 'functional MRI (fMRI)', 'structural MRI', and 'diffusion tensor imaging (DTI)'. These terms were used in various combinations and were searched for in all sections of the articles. Relevant references cited in papers found via this search were also reviewed. All data were extracted by four blinded reviewers (MS, MA, AC, MM). Our exclusion criteria encompassed several types of publications: editorials, letters to the editor, review articles and meta-analyses, case reports, postmortem studies, purely hypothetical papers, and any studies without an available English translation. Online search was performed in accordance with the PRISMA guidelines (Moher et al., 2009). The study period went from inception of databases to August 2023. The PRISMA flow diagram was presented in Fig. 1. Quality assessment of eligible papers was performed with the BIOCROSS evaluation tool (Wirsching et al., 2018). The tool includes 10 items covering 5 domains: ‘Study rationale’, ‘Design/Methods’, ‘Data analysis’, ‘Data interpretation’ and ‘Biomarker measurement’, aiming to assess different quality features of biomarker cross-sectional studies. Quality assessment of each included study was performed independently by three different authors (AM, LI, BA), and any disagreement was resolved by deliberation. We extracted the following data from eligible publications: 1) the number of participants; 2) clinical characteristics of patients; 3) cytokine or CRP levels (expressed as mean ± SD) with corresponding p-values for comparisons between patients and controls; 4) type of biological material used for assessment of cytokine or CRP levels (serum/ plasma); 5) information about tools used for assessment of cognitive performance; 6) correlation coefficients with corresponding p-values for associations between CRP or cytokine levels and cognitive performance; 7) imaging modalities descriptions. Initially, 1729 articles were identified through PubMed, Google Scholar, Scopus, and Web of Science searches. 453 duplicates were removed leaving 1276 unique articles for title and abstract screening (figure 1). Of the 1276 unique articles, 1052 were excluded via title and abstract screening for being review articles, animal studies, neurological disorders, and mental disorders other than BD, leaving 224 studies for full-text review. Of these 224 studies, 165 were excluded at full-text review as duplicates, not using cognitive measures, or absence of inflammatory markers, leaving 59 studies for quality assessment and data extraction.

Figure 1.

Flow Diagram of screening

Results

We identified 39 published clinical studies exploring the association of pro-inflammatory cytokines markers with cognitive performances in individuals diagnosed with BD. All studies were observational in nature and did not involve any anti-inflammatory and/or antioxidant treatments (table 1).

Table 1.

Summary of studies that explored changes in inflammatory biomarkers and cognitive measures in individuals with bipolar disorder (BD)

Study	Subjects Description (diagnosis, gender M/F)	Age in years (mean ± SD)	Inflammatory Biomarkers	Illness Definition	Cognitive Measures	Main Findings
Barbosa et al. (2012) Journal of Affective Disorders	25 euthymic BD type I (17F/8M) 25 HC (14F/11M)	BD 50.88 ± 9.11 HC 48.04 ± 7.08	BDNF TNF-α sTNFR2	Mini International Neuropsychiatric Interview (M.I.N.I-Plus)	Mini-Mental State Examination (MMSE) Frontal Assessment Battery (FAB)	BD patients had an impairment in executive functioning, particularly sensitivity of interference and inhibitory control in comparison with HC. Plasma levels of TNF-α were correlated with inhibitory control in BD patients while motor programming was negatively correlated with sTNFR2 plasma levels in controls.
Barbosa et al. (2018) Journal of Psychiatric Research	20 euthymic BD type I (15F/5M) 25 HC (16F/9M)	BD 43.8 ± 10.87 HC 43.52 ± 11.72	-IL-2 -IL-4 -IL-6 -IL-10 -INF-γ -TNF-α -IL-17A	Mini International Neuropsychiatric Interview (M.I.N.I)	Mini-Mental State Examination (MMSE) Brief Assessment of Cognition in Affective Disorders (BAC-A)	BD had higher plasma levels of sTNFR1, TNFα, IFN, IL2, IL4, IL6, IL10, and IL17 and lower cognitive performance in affective processing, verbal memory, working verbal memory, and executive functioning than HC. IL6 levels negatively correlated with global cognitive performance.
Barbosa et al. (2011) European Archives of Psychiatry and Clinical Neuroscience	53 BD type I 34 in mania (21F/13M) 19 in euthymia (11F/8M) HC 38 (20F/18M)	BD mania 49.6 ± 14.2 euthymia 44.5 ± 10.9 HC 42.9 ± 9.7	-TNF-α -sTNFR1 -sTNFR2	Mini International Neuropsychiatric Interview (M.I.N.I-Plus)		BD in mania had higher levels of sTNFR1 than BD in euthymia and controls.
Congio et al. (2022) Journal of Psychiatric Research	42 euthymic BD (31 type I, 11 type II) 33F/9M 27 MDD (23F/4M) 40 HC (23F/7M)	BD 39.69 ±11.28 MDD 48.74 ±9.97 HC 43.95 ±11.9	hsCRP	Structured clinical interview Childhood Trauma Questionnaire (CTQ)	Trail Making Test (TMT- A and B) Stroop test Phonemic and semantic verbal fluency tests	High levels of hs-CRP (≥ 5 mg/L) were linked with a history of childhood sexual and physical abuse, and worse neurocognitive performance.
Chakrabarty et al. (2019) Journal of Affective Disorders	51 euthymic BD type I (27F/24M) 20 HC (11F/9M)	BD 23.16 ±4.58 HC: 24.86 ±5.28	-IL-4 -IL-10 -IL-1-α -IL-6 TNF-α	Mini International Neuropsychiatric Interview (M.I.N.I)	Subtests from the Cambridge Neuropsychological Test Automated Battery (CANTAB)	TNFα was elevated at trend level significance in BD patients compared to HC, and was negatively associated with global cognition, processing speed, and working memory in patients.
M.H. Chen et al. (2021) CNS Spectrums	59 euthymic BD type I (36F/23M) 51 MMD (38F/13M) 51 HC (31F/20 M)	BD 37.24 ±9.56 MMD 37.16 ±12.04 HC 36.06 ±9.45	-IL-6 -TNF- α -CRP	Structured clinical interview	Wisconsin Card Sorting Test (WCST)	Patients with BD or MMD had higher CRP and TNF-α levels, and greater executive dysfunction in the WCST than the controls. TNF-α level was significantly associated with cognitive function deficits.
M.H. Chen et al. (2020) Journal of Affective Disorders	22 BD (first episode) (18F/4M) 22 MMD (first episode) (18F/4M) 22 HC (18F/4M)	BD 28.1 ±12.0 MDD 28.7 ±13.3 HC 27.4 ±10.2	sIL-6R CRP TNFR1	Clinical interview	-WCST	BD exhibited higher TNFR1 levels, more deficits in executive function and smaller grey matter (GM) volume in the middle frontal cortex (MFC) compared with MMD and HC. Correlation analysis revealed a positive association between the TNFR1 level and left MFC volume.
M.H. Chen et al. (2023) CNS Spectrums	35 BD (first episode) (24F/11M) 26 SCZ (first episode) (15F/11M) 29 MDD (first episode) (24F/5M) 22 HC (14F/ 8M)	BD 17.06 ±1.59 SCZ 17.92 ±2.59 MDD 17.07 ±0.8 HC 17±1.2	IL-2 IL-6 TNF-α CRP	Mini International Neuropsychiatric Interview (M.I.N.I)	Wisconsin Carding Sorting Test (WCST) Working memory task	BD had higher pro-inflammatory cytokines (IL-6, TNF-a, and CRP) and lower anti-inflammatory cytokines (IL-2) compared with SCZ, MDD and HC, and lower cognitive performance than HC. CRP levels were negatively associated with WCST scores.
P.S. Chen et al. (2022) International Journal of Immunopathology and Pharmacology	191 BD type II (major depressive status) (95F/96M) 112 HC (56 F/56M)	BD 32.24 ± 11.68 HC 33.04 ± 11.64	hsCRP	Schedule of Affective Disorder and Schizophrenia- Life Time	Continuous Performance Test (CPT) Finger-tapping test (FTT)	BD had higher CRP levels and worse cognitive function than the HC. The genotypes of CRP polymorphisms and interactions between the polymorphisms of CRP and life event changes had a significant influence on cognitive function
Chung et al. (2013) Neuropsychobiology	17 euthymic BD type I (10F/7M)	BD 31.3 ± 7.0	hsCRP	Structured Clinical Interview	MRI Wisconsin Card Sorting Test (WCST)	The left and right orbitofrontal cortex had a significantly negative relationship with serum hs-CRP levels. Analysis revealed a significant correlation between the left inferior orbitofrontal cortex and the indices of WCST performance.
Dickerson et al. (2013) Journal of Affective Disorders	107 BD 93 BD type I 14 BD type II (31M/76 F) 210 HC - Information not available	BD 36.3 ±13.4 HC -Information not available	hsCRP	Structured Clinical Interview for Diagnosis (SCID)	Repeatable Battery for The Assessment of Neuro-psychological status (RBANS) Wechsler Adult Intelligence Scale (WAIS III), Trail Making Test (TMT)	Lower RBANS scores in the High CRP group compared with the Low CRP group. There was an inverse relationship between CRP levels and performance on RBANS total, RBANS immediate memory, RBANS attention, RBANS language and Trail Making A.
Doganavsargil-Baysal et al. (2013) Human Psycho-pharmacology	54 BD euthymic (36F/18M) HC (13F/5M)	BD 39.46 ±11.62 HC 38.33 ±10.80	TNF-α sTNFR1 sTNFR2	Semi-Structured Clinical Interview	Wisconsin Card Sorting Test (WCST) Rey’s Auditory Verbal Learning Test (RAVLT)	BD had higher levels of sTNFr1 and sTNFr2 and lower cognitive performance on WCST and RAVLT than HC. TNF-a level was found to have a negative correlation with the delayed recall in RAVLT.
Garés-Caballer et al. (2022) Frontiers in Neurology	165 with severe mental illness (SMI) 42 BD type I (21F/11M) 30 SCZ (7F/23M) 35 MMD (24F/11M) 39 Diabetes Mellitus type 2 (T2DM) (9F/21M) 28 HC (18F/10M)	BD 50.0 ±9.5 SCZ 40.08 ±10.07 MMD 47.3 ±11.8 T2DM 57.3 ±9.3 HC 36.6 ±14.5	TNF-α IL-6 IL-10 CRP	Clinical Interview	Cognitive flexibility (e.g., Stroop Color and Word Test) Verbal Fluency Tasks Working Memory (e.g., Trail Making Test-B) Processing speed (e.g., Finger Tapping Test)	Participants with SMI and T2DM showed increased peripheral levels of IL-10 and TNF-α compared with HC. Individuals with BD and SZ had worse executive functioning compared to the other groups. In participants with BD executive functioning, variance was explained by a combination of pro-inflammatory factors (IL-6 and CRP).
Hamdani et al. (2015) Journal of Affective Disorders	42 euthymic BD type I and II with high seropositivity for T. Gondii (20F/22M) 36 HC (24F/12M)	BD 46.1 ±13 HC 38.1 ±15.2	IL6-mRNA	Diagnostic Interview for Genetic Studies” (DIGS)	California Verbal Learning Test (CVLT) Vocabulary subtest of the Wechsler Adult Intelligence Scale (WAIS III) A cognitive deterioration index (DI) was calculated with subtests extracted from the WAIS III	BD had higher seropositivity for T. gondii and higher levels of IL-6 mRNA expression than controls and performed poorly on neuropsychological measures compared with HC. The cognitive deterioration index (DI) was higher in BD and correlated to high IL-6 mRNA expression only among those infected by T. gondii.
Hope et al. (2015) Schizophrenia Research	111 BD type I/II (51F/60M) 121 SCZ (66F/55M) 241 HC (147F/89M)	BD 33 ±10 SCZ 36 ±10 HC 36 ±12	sTNF-R1 IL-1RA I-L6 hsCRP sCD40L OPG vWf	Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I)	Subtests for verbal cognition (Similarities and Vocabulary) and for performance abilities (Block design and Matrix reasoning) from Wechsler Abbreviated Scale of Intelligence (WASI)	BD and SCZ had higher levels of sTNF-R1, vWf, OPG and lower cognitive abilities than HC. In BD there was a significant association between general cognitive abilities and IL-1Ra and sCD40L.
Hoseth et al. (2016) Acta Psychiatrica Scandinava	117 BD type I/II and not otherwise specified. (71F/46M) 109 SCZ (50F/59M) 236 HC (132F/104M)	BD 32 ±21 SCZ 30 ± 13 HC 35 ± 17	sTNF-R1 OPG IL-1RA I-L6 SCD40L CRP vWf	Structured Clinical Interview for DSM-IV Axis I Disorders (SCIDI)	Wechsler Memory Scale (WMS-III): Logical Memory immediate (LM learning) and delayed recall (LM recall) California Verbal Learning Test (CVLT): learning and delayed free recall	Patients showed higher levels of sTNF-R1 and vWf compared to HC. There was a significant negative association between sTNF-R1 and verbal memory (LM learning and LM recall).
Hoseth et al. (2017) Schizophrenia Bulletin	247 BD (144F/103M) 569 SCZ (229F/340M) 624 HC (294F/330 M)	BD 32 ±18 SCZ 27 ±15 HC 32 ±13	TNF-molecules sTNFR1 sTNFR2	Structured Clinical Interview for DSMIV Axis I Disorders (SCID-I)	Digit Span Test— backward (Wechsler Adult Intelligence Scale (WAIS-III) Letter number sequencing (WAIS-III) Working Memory- Mental Arithmetic (WM-MA) Test	Both the patient group as a whole and the separate SCZ and BD groups had higher levels of TNF, sTNFR1, and sTNFR2, as well as the TNF/sTNFRs ratio compared to HC. There was a significant association between TNF/sTNFRs ratio and lower working memory scores, particularly stronger in the SCZ group compared to HC and BD.
Hua et al. (2021) Journal of Affective Disorders	52 euthymic BD type I (34F/24M) 27 euthymic BD type II (19F/8M) 51 HC (31F/20M)	BD I 38.31±9.12 BD II 37.33±11.86 HC 36.06±9.45	sIL-6R CRP sTNFR1	Clinical Interview	Word List Memory Task (WLMT) from Wechsler Memory Scale (WMS-III) Wisconsin Card Sorting Test (WCST)	Patients with BD I showed significantly higher levels of CRP, sIL-6R, and sTNFR1 than HC. BD I/II demonstrated poorer performance in WLMT and WCST compared to HC. sTNFR1 levels were negatively correlated with cognitive function only in BD I.
K. L. Huang M.H et al. (2022) Psychiatry and Clinical Neurosciences	37 BD I (19F/18M) 33 BD II (20F/13M) 25 MDD (18F/7M) 54 HC (34F/20M)	BD I 34.9±8.2 BD II 31.5±8.5 MDD 34±10.1 HC 32.1±7.7	sIL-6 R CRP sTNF-αR1	Mini International Neuropsychiatric Interview (M.I.N.I)	Word List Memory Task (WLMT) Wisconsin Card Sorting Task (WCST) 2-back task Go/No-Go task. Divided Attention Task	Patients with BD type I had higher levels of sTNF-αR1 than patients with MDD and HC. Patients with BD type I performed worse on WLMT, WCST, 2-back task, and a divided attention task than the other groups. sTNF-αR1 levels were negatively correlated with cognitive function (WLMT and divided attention_auditory).
K. L. Huang et al. (2022) CNS Spectrums	38 BD (Recurrent) (19F/19M) 31 BD (First-episode) (24F/7M) 43 HC (27F/16M)	BD Recurrent 28.63 ±3.82 BD First-episode 25.26±5.39 HC 26.16±4.04	IL-6 TNF-α CRP	Clinical Interview	Wisconsin Card Sorting Test (WCST)	Patients with BD had significantly higher TNF-α compared with the HC. Patients in the multiple-episode BD group performed significantly worse in the WCST and lower levels of CRP than those in the first-episode BD group. No significant associations of CRP and TNF-α levels with executive function were observed.
Johnsen et al. (2016) BMC Psychiatry	124 (11 BD) patients with Severe Mental Illness (SMI) (84M/40F)	SMI 33.5±12.4	CRP	Clinical Interview	Repeatable Battery for the Assessment of Neuro-psychological Status (RBANS)	There was a significant inverse relationship between overall cognitive performance and CRP level at baseline. At follow-up (the mean interval between baseline and follow-up was 28.3 (11.1) days) the association between CRP level and cognition was no longer present.
Kauer-Sant'Anna et al. (2009) International Journal of Neuropsycho-pharmacology	60 BD I patients: 30 BD early-stage (13M/17F) 30 BD late stage (9M/21F) 60 HC: 30 HC group 1 (10M/20F) 30 HC group 2 (11M/19F)	BD early stage 22.4 ±3.9 BD late stage 41.4 ± 8.4 HC group 1 22.1±3.6 HC group 2 43.2±6.4	TNF-α IL-6 IL-10 BDNF	Mini International Neuropsychiatric Interview (M.I.N.I) Structured Clinical Interview for DSM-IV -Axis I Disorders (SCID-I)		TNF-a and IL_6 levels were significantly higher in both early-stage and late-stage BD patients compared to controls, while IL-10 was increased only in early-stage BD, not in late-stage BD. BDNF levels decreased significantly in BD late stage.
King et al. (2019) Frontiers Psychiatry	15 BD II (12 euthymic and 3 depressed) -Information not available 13 HC -Information not available	BD /HC: 22-57 -Information not available	IL-6 TNF-α CRP	Clinical interview	Working Memory (e.g., Trail Making Test-B) Rey’s Auditory Verbal Learning Test (RAVLT) Digit Symbol Substitution test - Wechsler Adult Intelligence Scale (WAIS-III) The Weschler test of Adult Reading	There was no significant difference between glutamate levels and cognitive test scores. Worse performance on the Digit Symbol Task was associated with higher depressive severity. There were no significant differences in TNFa, IL-6, or CRP levels between BP-II and healthy controls.
Lizano et al. (2021) Molecular Psychiatry	SCZ 79 Psychotic BD 61 (PZ 60M/80 F) 60 HC (26M/34F)	PZ 34.3± 13.3 HC 37.6± 13.5	IL-1α /IL-1β IL-2 IL-4 IL-6 IL-8 IL-10 IL-23p40 IL-12p70 IFN-γ TNF- α / TNF-β CRP Flt-1 VEGF (C/D)	Clinical Interview	Brief Assessment of Cognition in Schizophrenia (BACS) Spatial Span subtest from Wechsler Memory Scale (WMS) Dot pattern expectancy task reaction time (DPX RT) Antisaccade error rate(ER)	Elevated levels of specific markers, including TNFα, IL6, VEGF and CRP, were associated with cognitive and neuroanatomical differences. Notably, higher IL6 and CRP levels were linked to poorer cognition in probands, while elevated CRP and VEGF levels were associated with increased antisaccade errors, but no such correlations were found in healthy controls.
Lotrich et al. (2014) International Journal of Geriatric Psychiatry	21 BD euthymic (BD-I and BD-II) (8M/13F) 26 HC (12M/14F)	BD 64.8±9.1 HC 65.5±8.4	IL-1RA IL-6 TNF- α Eotaxin BDNF	Structured Clinical Interview for axis I DSM-IV (SCID-IV)	21 individual test and Four zeta-scores (language, delayed memory, visuomotor ability and information processing speed/executive function). -No other information available	Elevated levels of the inflammatory cytokine IL-1RA were found in older adults with BD and were associated with cognitive deficits. The most pronounced effects were observed in processing speed, executive function, and cognitive control tests.
Millett et al. (2021) Molecular Psychiatry	222 BD Euthymic 179 BDI 43 BDII (105F/117M) 52 HC (30F/22M)	BD 44±11.9 HC 39±13.4	Serum CRP	Structured Clinical interview for the DSM-V (SCID-V)	Matrics consensus cognitive battery (MCCB) California verbal learning test (CVLT) Wechsler’s Abbreviated scale for intelligence (WASI) Wide Range achievement test-3 (WRAT-3) Controlled oral word association task (COWAT) Wisconsin card sorting task (WCST)	Participants with CRP levels ≥ 5 mg/L exhibited poorer performance on various measures of executive functioning, MATRICS processing speed, and MATRICS reasoning and problem-solving skills compared to those with lower CRP levels. CRP was identified as a significant positive predictor of cognitive decline proxies.
Millett et al. (2020) Brain, Behavior, and Immunity	219 euthymic BD (176 BDI and 43 BDII) (103 F/116M) 52 HC (30F/22M)	BD 43.6±11.9 HC 38.9±13.4	TNF-α sTNF-R1 sTNF-R2	Structured Clinical Interview for axis I DSM-IV (SCID-IV)	Controlled Oral Word Association Task (COWAT) Wisconsin Card Sorting Task (WCST) Matrics Consensus Cognitive Battery (MCCB) Stroop color and word test	Patients with BD demonstrated higher levels of sTNFr1 and sTNFr2 compared to HC. Individuals with BD exhibited lower performance on assessments like WCST and RAVLT in comparison to the control group. sTNF-R2 was also significantly associated with Stroop test results.
Milton et al. (2021) European Psychiatry	502 BDI (233F/269M) 467 BDII (208F/259M) 5963 MDD severe (3422F/2541M) 10739 MDD moderate (7321F/3418M) 5739 MDD single episode (3585F/2154M) HC 60.858 (29557F/31201M)	BDI 55±8 BDII 56±8 MDD severe 56±8 MDD moderate 56±8 MDD single episode 57±8 HC 58±8	Serum CRP	Structured Clinical Interview for axis I DSM-IV (SCID-IV)	Test of visuospatial memory (e.g. Reaction time and Pairs matching).	Higher CRP levels were linked to mood disorders, but there was a small positive connection between CRP and memory test performance. Mood disorders were consistently tied to poorer cognitive function, especially in reaction time and pairs matching tasks.
Mora et al. (2019) European Psychiatry	52 euthymic BD (34 BD I and 18 BD II) (26F/26M) 32 manic patients (31 BD I and 1 BD II) (14F/18M) 49 HC (28F/21M)	Euthymic BD 47.52±11.9 Manic BD 41.25±12.9 HC 48.3±12.1	IL-6 IL-10 TNF-α BDNF	Structured Clinical Interview for axis I Disorders (SCID-IV)	Vocabulary, Block Design and Digits Subtests (WAIS-III) Wisconsin Card Sorting test (WCST) Stroop Color and Word Test FAS verbal fluency task Trail Making Test (TMT- A and B) Conners’ continuous Performance Test II The California Verbal Learning Test (CVLT) Rey-Osterrieth Complex figure (RCFT)	BD patients had higher levels of IL-6 compared to healthy controls, IL-10 levels were not significantly different and BDNF levels were significantly lower. The study revealed that bipolar individuals had difficulties in various cognitive domains, including executive functioning and verbal memory.
Peters et al. (2022) Journal of Affective Disorders	119 BD (49F/70M)	47.49 ± 9.9	CRP	Structured Clinical Interview for DSM-V (SCID-5)	Affective Go/no go task (CANTAB)	Increased CRP levels was linked to affective inhibition performance in euthymic individuals with BD. BD patients had difficulties across various cognitive domains including executive functioning and verbal memory.
Poletti et al. (2021) Journal of Psychiatric Research	76 BD patients with a depressive episode (53F/23M)	BD 47.11± 11.32	IL-1β, IL-1ra, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-12, IL-13, IL-15, IL-17A IP-10 – CXCL10 MCP-1 -CCL2 MIP-1α - 1β TNF-α IFN-γ FGF PDGF-BB G-CSF GM-CSF VEGF	Clinical Interview	Brief Assessment of Cognition in Schizophrenia (BACS)	Elevated levels of specific interleukins, such as TNF-α (linked to verbal memory and information processing) and IL-6 (associated with executive functions and verbal fluency), are connected to increased risk of cognitive impairment. CRP levels were also linked to impaired affective inhibition in euthymic BD patients, suggesting a role for inflammation in cognitive-affective processing integration, particularly for negative stimuli in BD.
Rahmani et al. (2021) Journal of Affective Disorders	60 BD (34 symptomatic 22F/12M, 26 asymptomatic 15F/11M) 51 HC (25F/26M)	BD sympt. 17.45±1.59 BD asympt. 17.59±1.95 HC 17.15±1.38	CRP IL-1β IL-6 IL-10 IL-4 TNF-α	Schedule for Affective Disorders and Schizophrenia for School-Age Children-Present and Lifetime Version (KSADS- PL)	Wechsler Abbreviated Scale of Intelligence (WASI) Cambridge Neuropsychological tests Automated Battery (CANTAB) Intra/extradimensional shift (IED)	In symptomatic adolescents with BD, certain inflammatory markers (TNF, IL-1β, IL-6, IL-10, and CRP) were linked to neurocognitive flexibility. Higher TNF levels were associated with improved flexibility, while increasing IL-1β levels led to more errors during the executive function shift stage. Additionally, lower pro-inflammatory to anti-inflammatory ratios were linked to reduced flexibility, especially in symptomatic BD adolescents.
Rolstad et al. (2015) European Neuro-psychopharmacology	78 BD (48 BD I and 30 BD II) (47F/31M) 86 HC (40F/46M)	BD: 38.2 ± 13.4 HC ±37.8 13.4	CSF markers: YKL-40 TIMP-1 MPC-1 sCD14	Structured Clinical Interview for DSM-IV (SCID-IV) Mini International Neuropsychiatric interview (M.I.N.I.)	Wechsler’s Adult Intelligence Scale version III (WAIS-III) (e.g., Digit Symbol,Block Design) Delis-Kaplan executive function system (D-KEFS) (e.g., Verbal fluency Test, Color-Word interferance, Design Fluency, Trail Making Test number-letter sequencing, Tower Test, Continuous Performance Test)	Elevated levels of CSF neuroinflammatory marker (YKL-40) were significantly linked to executive function impairments in euthymic bipolar disorder individuals, while these associations were not observed in HC.
Sanchez-Autet et al. (2018) Journal of Affective Disorders	224 BD I, II NOS (146F/48M)	BD M 45.7 ± 13.6; F 47.8 ±11.8	Homocysteine (HCY) CRP	Structured Clinical Interview for DSM-IV Axis Disorders (SCID-I)	Screen of Cognitive Impairment (SCIP)	In men, higher levels of C-reactive protein (CRP) were associated with poorer performance on the FAST5, specifically in aspects related to interpersonal relationship functioning. In women, higher CRP levels were associated with lower cognitive performance, affecting SCIP_Traw score, immediate verbal learning (SCIP2), verbal fluency (SCIP3), and delayed verbal learning (SCIP4).
Tseng et al. (2020) Brain, Behavior, and Immunity	25 euthymic BD (16F/9M) on medication 43 HC (22F/21M)	BD 36.48± 12.69 HC 31.5±9.96	Hs-CRP	Mini International Neuropsychiatric Interview (MINI)	fMRI Wisconsin Card-Sorting test (WCST) Continuous Performance test (CPT)	BD patients had higher hs-CRP levels than controls. This elevated inflammation level was associated with increased dorsal corticostriatal connectivity, specifically from the dorsal striatum to the vlPFC, in BD patients. These altered connectivity patterns were linked to cognitive deficits in attention and executive function observed in BD patients.
Vai et al (2022) Brain, Behavior & Immunity-Health	84 BD (64 depressed, 16 maniac, 4 euthymic) (29F/55M)	BD 47.01 ± 11.85 Depressed 46.39 ±11.52 Maniac 52.31 ±11.41Euthymic 37.25±12.61	IL-6	Clinical Interview	MRI Cognitive Questionnaire (CQ)	In BD patients, a higher level of peripheral IL-6 was associated with lower GM volumes in supragenual ACC and reduced rs-FC between this area and medial OFC. Additionally, higher IL-6 levels were correlated with cognitive biases associated with depressive episodes.
Wiener et al. (2017) Journal of Neuroimmunology	48 BD drug-free (36F/12M) 48 MDD (36F/12M) 48 HC (36F/12M)	BD: 21.92 ±2.32 MDD 21.81 ±2.14 HC: 21.88 ±2.31	IL-6 IL-10	Mini International Neuropsychiatric Interview (MINI) Structured Clinical Interview for DSM-IV (SCID)		Bipolar disorder (BD) and major depressive disorder (MDD) subjects had similar peripheral IL-6 and IL-10 levels compared to controls, but BD and MDD individuals exhibited greater functional impairment. Notably, in BD patients, IL-6 and IL-10 levels correlated positively with functional impairment.
Zazula et al. (2022) The World Journal of Biological Psychiatry	31 BDI-II (25F/6M) 27 HC (18F/9M)	BD 39.5 ± 11.5 HC 38.7 ±13.7	TNF-α sTNF-R1 sTNF-R2	Portuguese version of the Structured Clinical Interview for the DSMIV (SCID-IV)	CogState Research Battery Groton Maze Learning Task (GML) Set-shift Task (SETS) Two-Back Task (TWOB)	In BD patients, sTNFR2 displayed positive correlations with the Groton Maze Learning Task and Set-Shifting Task, and weak negative correlations with the Two-Back Task. No significant correlations were observed between TNF-a and sTNFR2, or between TNF-a and any of the administered neurocognitive tasks.
Zhang et al. (2022) Journal of Psychiatric Research	96 UD (65F/31M) 111 BD (14 BD-I, 97 BD-II) (62F/49M)	UD 30.3±12.6 BD 28.6 ±11	TG TCh HDL-C LDL-C ApoA1 - ApoB Lpα IL-1β IL-6 CRP BDNF	Clinical Interview	Montreal Cognitive Assessment (MoCA)	BD patients displayed significantly lower cognitive scores, particularly in the language domain, indicating more severe cognitive impairments compared to UD. Elevated levels of ApoB and CRP, along with reduced BDNF levels, were associated with cognitive differences.

SCZ: Schizophrenia; MDD: Major Depressive Disorder; UD: Unipolar Depression; HC: Healthy Controls; PZ: Patients; SMI: Several mental illness; sTNF-R1: Tumor Necrosis Factor Receptor 1; sTNF-R2: Tumor Necrosis Factor Receptor 2; TNF-α / TNF-β: Tumor Necrosis Factor α/β; IL-1Ra: Interleukin Receptor Antagonist; IL-1α / IL-1β: Interleukin 1 α/β; IL-2: Interleukin 2; IL-4: Interleukin 4; IL-6: Interleukin 6;IL-7: Interleukin 7; IL-8: Interleukin 8; IL-9: Interleukin 9; IL-10: Interleukin 10; IL-12: Interleukin 12; IL-13: Interleukin 13; IL-15: Interleukin 15IL-17A: Interleukin 17 A; IL-23p40: Interleukin 23 p40; IL-12p70: Interleukin 12 p70; IFN-γ: Interferon γ; Flt-1: Fms-like Tyrosine Kinase 1; IP-10: Interferon gamma-induced protein 10 (CXCL10); MCP-1: Monocyte Chemoattractant Protein-1 (CCL2); MIP-1α and MIP-1β: Macrophage Inflammatory Protein-1α and 1β; FGF: Fibroblast Growth Factor; PDGF-BB: Platelet-Derived Growth Factor BB; G-CSF: Granulocyte Colony-Stimulating Factor; GM-CSF: Granulocyte-Macrophage Colony-Stimulating Factor; VEGF-C/ VEGF-D: Vascular Endothelial Growth Factor C/D; hsCRP: High Sensitivity C-Reactive Protein; OPG: Osteoprotegerin; vWF: von Willebrand Factor; sCD40L: Soluble Cluster of Differentiation 40 Ligand; BDNF: Brain-Derived Neurotrophic Factor; HCY: Homocysteine; TG: Triglycerides; TC: Total Cholesterol; HDL-C: High-Density Lipoprotein Cholesterol; LDL-C: Low-Density Lipoprotein Cholesterol; ApoA1: Apolipoprotein A1; ApoB: Apolipoprotein B; Lpα: Lipoprotein (a); T2DM: Type 2 Diabetes Mellitus; CSF: Cerebrospinal Fluid; YKL-40: Chitinase-3-like protein 1 (also known as YKL-40); TIMP-1: Tissue Inhibitor of Metalloproteinases-1;; sCD14: Soluble Cluster of Differentiation 14.

C-reactive Protein (CRP)

CRP is a pentameric protein, whose level rises in blood plasma in response to inflammation. CRP is an acute-phase reactant protein that is primarily induced by the IL-6 action on the gene responsible for the transcription of CRP during the acute phase of an inflammatory/infectious process (Nehring et al., 2022). The association between CRP levels and cognitive performance was tested in sixteen studies on BD patients (M. H. Chen et al., 2023; P. S. Chen et al., 2022; Congio et al., 2022; Dickerson et al., 2013; Hope et al., 2015; Hoseth et al., 2016; M. H. Huang et al., 2022; Johnsen et al., 2016; King et al., 2019; Lizano et al., 2021; Millett et al., 2021; Milton et al., 2021; Peters et al., 2022; Rahmani et al., 2021; Sanchez-Autet et al., 2018; Zhang et al., 2022). In three studies with healthy controls, CRP levels were not significantly higher in the patients' group (Hope et al., 2015; King et al., 2019; Millett et al., 2020). In the study by Dickerson et al. (2013) on patients with BD type I and II with mild symptoms, cognitive functioning was measured with the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) and two subtests from the Wechsler Adult Intelligence Scale III (WAIS-III): Information and Letter Number Sequencing, and Part A of the Trail Making Test. Global cognitive performance was significantly lower in patients with CRP levels higher than the 90th percentile and the 75th percentile of the control group. There were negative correlations between CRP levels and the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) total score, immediate memory score, attention score, learning score and Trail Making Test, part A (TMT-A) score. Consistent with the results of Dickerson et al. (2013), Millet et al. (2021), who conducted a study in a large cohort (N=222) of euthymic patients with BD (BD I and BD II), reported worse performance on several measures of executive functioning (Wisconsin Card Sorting Test (WCST); Stroop color), MATRICS processing speed, reasoning and problem solving and Reading Mind in the Eyes in participants with high levels of CRP (≥ 5 mg/L) relative to those with lower CRP. In a longitudinal study Johnsen et al. (2016) examined acutely relapsed patients with psychosis including patients with BD. The authors reported that CRP levels negatively correlated with scores of attentions, delayed memory and global cognitive performance on admission but not after 4 weeks of treatment. Alternatively, to studies which have established the relationship between individual peripheral inflammatory markers and cognition, Lizano et al. (2021) assessed 15 peripheral inflammatory markers in clinically stable psychosis probands with schizophrenia spectrum disorder and psychotic bipolar disorder, and matched healthy controls. Levels of IL6, TNFα, Vascular endothelial growth factor (VEGF), and CRP were significantly higher in psychosis probands compared to healthy controls. Higher IL6 and CRP levels were associated with poorer cognition (Brief Assessment of Cognition in Schizophrenia (BACS) total score), while higher levels of CRP and VEGF were associated with worse inhibitory behavioral control (i.e., elevated antisaccade error rate) in probands. Peters et al. (2022) conducted a study to assess the relationship between the inflammatory marker C-reactive protein (CRP) and affective inhibition performance in individuals with BD. Participants completed the Affective Go/No-Go Task. The authors found that higher levels of CRP were associated with a reduced ability to discriminate negative words compared to positive and neutral target conditions, and faster response times for both negative hits and commissions, suggesting that inflammation may have a significant impact on inhibitory control when attention is focused on negative stimuli. Recently, Zhang et al. (2022) compared cognitive impairments with inflammatory biomarkers in unipolar and bipolar depression. The authors found that CRP level in bipolar depression was higher than unipolar depression, and negative associated with cognitive ability measured with Montreal Cognitive Assessment (MoCA). More recently, M.H. Chen et al. (2023) explored the complex relationship of pro- and anti-inflammatory cytokines with cognitive function in adolescents with first-episode schizophrenia, bipolar disorder, major depressive disorder, and sex-matched controls. Patients with bipolar disorder were found to exhibit the highest levels of CRP, IL-6, and TNF-α among the four groups. The authors observed that among the assessed inflammatory markers, only CRP levels were negatively associated with measures of executive functioning. In another study, K.L. Huang et al. (2022) compared patients with first episode and multiple-episode bipolar disorder. The findings revealed that patients with multiple-episode BD exhibited lower CRP levels, and more impaired executive function evaluated using the WCST, compared to those with first episode, irrespective of demographic and clinical factors, which may suggest the potential role of CRP levels in the clinical stages of BD. In a large cohort study (Milton et al., 2021) was investigated the magnitude of associations between CRP, lifetime history of bipolar disorder type I and type II or major depression, and cognitive functions. 84,268 UK Biobank participants were involved in this important study. Participants were also categorized into quintiles based on their serum CRP level with one representing the lowest serum CRP levels and five representing the highest serum CRP levels. While increased serum CRP levels were associated with mood disorder history, the highest serum CRP concentrations were significantly correlated with lower cognitive performance on a reaction time test.

Despite these positive findings others failed to demonstrate an association between peripheral levels of CRP and cognition in patients with BD, in five studies of bipolar disorder type I, no significant correlations between CRP levels and cognitive performance on tasks measuring verbal cognition (Hope et al., 2015), visuospatial memory (Milton et al., 2021), neurocognitive flexibility (Rahmani et al., 2021), verbal memory (Hoseth et al., 2016) and executive functioning (M.H. Huang et al., 2022). Another study revealed no significant correlations between CRP levels and cognitive performance in euthymic patients with bipolar disorder type II on executive function (Trail Making Test-part B; TMT-B), processing speed (Digit Symbol Task; DST) and verbal learning (Rey Auditory Verbal Learning Test; RAVLT) (King et al., 2019). Variables which may account for these discrepancies may be related to the influence of gender differences, genetic and environmental factors. In the study by Milton et al. (2021) higher CRP levels were associated with mood disorder history suggesting its potential role in the clinical stages of BD. Moreover, Sanchez-Autet et al. (2018) suggested gender difference in biomarker modulation of cognition and psychosocial functioning in BD; while in men high CRP levels correlated with poor real-world functioning (interpersonal relationships and financial management, respectively), in women high CRP levels correlated with cognitive dysfunction (global cognitive performance, verbal fluency, and immediate and delayed verbal learning dimensions). The study of Congio et al. (2022) aimed to investigate the relationship between inflammation, cognitive impairment, childhood trauma, and physical activity, in patients with bipolar disorder, depressive disorder, and healthy controls. Cognitive functioning was assessed using the Trail Making Test Parts A and B, Stroop test, Symbol Digit Substitution (SDS) test, and phonemic and semantic verbal fluency tests. Patients exhibited poorer cognitive performance (i.e., lower scores in verbal fluency tasks, encompassing both phonetic and semantic aspects, Stroop color-word cards, and Trail Making Test Part A (TMT-A) and Part B (TMT-B) compared to the control group. Particularly, the results indicate that higher levels of high-sensitivity C-reactive protein (hs-CRP) were associated with a history of childhood sexual and physical abuse, as well as worse neurocognitive performance in BD. Recently, P.S. Chen et al. (2022) showed a significant association between the polymorphisms of CRP and cognitive function. The results in drug-naïve depressed BD patients showed that the genotypes of CRP rs1800947 were significantly associated with the Finger-tapping test (FTT) and the Continuous Performance Test (CPT) tests. BD patients with the CC genotype of CRP rs1800947 had better cognitive performance than G carriers. In addition, the genotypes of CRP rs1205 and rs2794520 were significantly associated with performance on the FTT tests. BD patients with the CC genotypes of rs1205 and rs2794520 had better cognitive performance on the nondominant hand FTT than those with T carriers. Furthermore, the study suggested that the interactions between the polymorphisms of CRP and life event changes had a significant influence on cognitive function in BD patients after receiving VPA treatment for 12 weeks.

Tumor Necrosis Factor-alpha (TNF- α), soluble and -2 (tumour sTNFR-necrosis 1/-2) factor receptors-1

Tumor Necrosis Factor-alpha (TNF-α) is an inflammatory cytokine produced by macrophages/ monocytes during acute inflammation and is responsible for a diverse range of signalling events within cells, leading to necrosis or apoptosis. TNF- α shows its activity through soluble tumour necrosis factor receptors-1 and -2 (sTNFR1, sTNFR2) (Gough & Myles, 2020). The association between TNF-α and/or its receptors levels and cognition was assessed in many studies from patients with BD (Barbosa et al., 2012, 2018; Chakrabarty et al., 2019; M.H. Chen et al., 2020, 2021; Doganavsargil-Baysal et al., 2013; Garés-Caballer et al., 2022; Hope et al., 2015; Hoseth et al., 2016; Hua et al., 2021; K. L. Huang et al., 2022; M.H. Huang et al., 2022; King et al., 2019; Millett et al., 2020; Mora et al., 2019; Zazula et al., 2022). Barbosa et al. (2012) investigated executive functioning through the Frontal Assessment Battery (FAB) and plasma levels of inflammatory markers (i.e., TNF-α, sTNFR1, sTNFR2) in a sample of BD type I euthymic patients, as well as age- and gender-matched healthy controls. The FAB consisted of six sub-tests that explore neurocognitive processes related to the frontal lobes: conceptualization; mental flexibility: motor programming; sensitivity of interference; inhibitory control; and environmental autonomy. The authors found that there were no significant differences in plasma levels of TNF-α, sTNFR1, and sTNFR2 between BD patients and healthy controls. Plasma levels of TNF-α were positively correlated with inhibitory control in BD patients, while motor programming was negatively correlated with sTNFR2 plasma levels in controls. In a subsequent study by Barbosa et al. (2018), it was found that patients with remitted bipolar disorder (BD) type I exhibited elevated plasma levels of various inflammatory markers, such as sTNFR1, TNFα, IFN, IL2, IL4, IL6, IL10, and IL17, compared to controls. However, the study did not demonstrate a direct link between TNF-α or sTNFR1 levels and cognitive impairment. Cognitive function was evaluated using the Brief Assessment of Cognition in Affective Disorders (BAC-A), which comprises eight tasks. These tasks are designed to assess different cognitive domains: affective interference, verbal memory, working memory, motor speed, combined attention and motor speed, semantic and letter fluency, as well as executive functioning. In another study conducted by Doganavsargil-Baysal et al. (2013), a significant difference was found between euthymic BD type I patients and controls in terms of sTNFR1 and sTNFR2 serum levels, whereas no difference was detected between the BD group and the control group in levels of TNF-α. The Wisconsin Card Sorting Test (WCST) and Rey’s Auditory Verbal Learning Test (RAVLT) were applied to both patient and control groups to evaluate cognitive functions. Even though the concentration of TNF-α in BD patients was not different from controls, it correlated negatively with RAVLT-delayed recall, while there was no relation between both sTNFR1 and sTNFR2 and delayed recall. The authors argued that one possible explanation for this discrepancy is that TNF-α simultaneously causes the upregulation of AMPA receptors by the mediation of sTNFR1 and sTNFR2, and a decrease in synaptic inhibition of gamma-aminobutyric acid (GABA) A receptors. This dual action clarifies the absence of correlation between sTNFR1, sTNFR2, and delayed recall. No significant correlation was found between the serum levels of TNF-a, sTNFR1, and sTNFR2 and the results of the WCST in both study groups. The study conducted by Hope et al. (2015) enrolled patients with schizophrenia spectrum (i.e., schizophrenia, schizophreniform and schizoaffective disorder) and bipolar patients bipolar spectrum disorder (i.e., bipolar I disorder; bipolar II disorder, BD not otherwise specified) and healthy controls. Serum concentrations of sTNF-R1, interleukin 1 receptor antagonist (IL-1Ra), osteoprotegerin (OPG), vonWillebrand factor (vWf), CRP, interleukin-6 and CD40 ligand were measured. General intellectual abilities were assessed using the Wechsler Abbreviated Scale of Intelligence (WASI) which includes two subtests for verbal cognition (Similarities and Vocabulary) and two tests for performance abilities (Block design and Matrix reasoning). The findings revealed main effects of diagnosis on the following inflammatory markers: sTNF-R1, vWf, and OPG, indicating increased levels in both groups of patients compared to healthy controls, whereas no significant differences were observed between bipolar and schizophrenia patients in any of the inflammatory markers studied. Regarding the association between inflammatory markers and cognitive functioning, the authors found only significant negative associations between sCD40L and IL-1Ra serum levels, and general cognitive abilities in bipolar disorder patients (see below). In a comparable study that encompassed patients with non-euthymic bipolar spectrum disorders, individuals with schizophrenia spectrum disorders, and healthy control subjects, verbal memory was evaluated. The assessment tools used were the Wechsler Memory Scale-Third Edition (WMS-III), focusing on Logical Memory for immediate (LM learning) and delayed recall (LM recall), and the California Verbal Learning Test, which measured summed recall across learning trials (CVLT learning) and delayed free recall (CVLT recall). The main finding of the study was a negative association between sTNF-R1, and performance on LM learning and LM recall tests in all participants (including patients and controls). Hoseth et al. (2016) found a moderate negative association between sTNF-R1, and performance on verbal memory learning and recall tests as measured by the WMS-III Logical Memory in both patients with psychotic disorders (including bipolar disorder) and in healthy controls. In another study, M.H. Chen et al. (2020) found a significant association between TNFR1 levels and executive function measures (WCST). In the study conducted by Millet et al. (2020), euthymic patients with BD-I and BD-II, and healthy controls did not differ on serum levels of TNF family molecules (TNF-α, sTNF-R1, sTNF-R2). Cognitive functioning was measured using the Controlled Oral Word Association Task (COWAT, the Stroop color and word test, the Wisconsin Card Sorting Task (WCST) and the MATRICS Consensus Cognitive Battery (MCCB; including 10 standardized cognitive tests with measures in 7 cognitive domains: speed of processing attention/vigilance, working memory, verbal learning, visual learning, reasoning and problem solving, and social cognition). A series of correlation analyses revealed significant negative associations between TNF-α and measures of executive function, such as the Stroop color-word test and the WCST. sTNF-R2 was also significantly negatively associated with Stroop test. The relationship between inflammatory markers and cognition seemed to be specific to executive functioning given that the authors did not find a significant correlation between TNF molecules and the neurocognitive scores from the MCCB. Interestingly, the authors also found that higher sTNF-R1 concentrations correlate with ‘late-stage’ BD illness, as opposed to those in early-stage illness. Moreover, circulating levels of TNF molecules mediate the relationship between the number of severe mood episodes and executive functioning, suggesting that TNF family molecules may play an important role in the pathophysiology of neuroprogression in BD. Chakrabarty et al. (2019) evaluated the relationship between serum levels of pro-inflammatory cytokines (TNF-α, IL-6 and IL-1 α), anti-inflammatory cytokines (IL-4 and IL-10) and cognitive functioning in clinically stable patients with BD type I in the early stage of the disease (i.e., first manic or mixed episode in the 3 months prior to enrolment), with the aim of identifying inflammatory mediators that may significantly impact cognitive functioning early in the disease course. Subtests of the Cambridge Neuropsychological Test Automated Battery (CANTAB), as well as additional measures of verbal memory, attention, and working memory, were used to assess cognitive functions, namely: i) processing speed; ii) verbal memory; iii) nonverbal memory (e.g., spatial recognition memory (SRM); iv) working memory; v) executive functioning. In this study, TNF-α was elevated at trend level significance in BD patients compared to healthy controls and was negatively associated with global cognition, processing speed, and working memory in patients. More recently, Zazula et al. (2022) examined cognition-immune interactions between serum levels of TNF-a, sTNFR1 and sTNFR2 and specific neuropsychological domains of cognitive functioning (executive function, working memory) in euthymic patients with a diagnosis of Type I or Type II BD in comparison with controls. Three tasks from the Cogstate Research Battery were used to assess executive functions (i.e., Groton Maze Learning Task (GML), Set-Shifting Task (SETS)) and working memory (i.e., Two-Back Task (TWOB)). Peripheral TNF-a levels were significantly different between BD and control groups. A trend towards significance was found for sTNFR1, while no differences between groups was found for sTNFR2. Among BD participants, there were significant correlations between sTNFR2 and neurocognitive tasks evaluating both executive function and working memory. Hua et al. (2021) enrolled patients with remitted BD type I/II and healthy controls. Pro-inflammatory cytokines, including soluble interleukin-6 receptor (sIL-6R), C-reactive protein, and soluble tumour necrosis factor receptor 1 (sTNFR1) were measured, and performance in the Word List Memory Task (WLMT) and Wisconsin Card Sorting Task (WCST) was assessed. Significantly elevated levels of sTNFR1 were observed among patients compared with controls. Furthermore, sTNFR1 levels were negatively correlated with cognitive function measured using the WLMT and WCST. Recently, the study conducted by M. H. Huang et al. (2022) revealed that patients with bipolar disorder type I have higher levels of the TNF-R1 compared to those with major depressive disorder (MDD) and healthy controls. Furthermore, TNFR1 levels were negatively correlated with cognitive impairment measured using the Word List Memory Task (WLMT) and divided attention. Garés-Caballer et al. (2022) evaluated the associations between peripheral inflammatory biomarkers and cognition and social functioning across chronic somatic and mental illnesses from a transdiagnostic perspective. Impairments in cognitive flexibility, verbal fluency, working memory and processing speed in participants with BD were correlated with CRP, IL-1RA, IL-6, and TNF-α. Similar results were reported by M.H. Chen et al. 2023, who investigated cognitive function of adolescents with first-episode schizophrenia, bipolar disorder, or major depressive disorder. The findings revealed that adolescents with bipolar disorder had higher levels of CRP, IL-6, and TNF-α but lower levels of IL-2 compared to the other groups. Those with schizophrenia showed the most significant impairments in cognitive function based on the Wisconsin Card Sorting Test (WCST) and a working memory task. Notably, only CRP levels were negatively correlated with WCST scores. While these studies suggest a moderate/strong correlation between TNF molecules and cognitive functioning, other studies on patients with BD did not report any correlation of TNF-α and its receptors with cognitive functioning (Barbosa et al., 2018; M. H. Chen et al., 2023; Garés-Caballer et al., 2022; K. L. Huang et al., 2022; King et al., 2019; Mora et al., 2019). Notwithstanding, several studies on BD reported that levels of TNF molecules showed a positive correlation with the number of episodes and illness duration, suggesting that enhanced inflammation is associated with disease progression (Barbosa et al., 2011, 2014; Doganavsargil-Baysal et al., 2013; K. L. Huang et al., 2022; Kauer-Sant’Anna et al., 2009).

IL-6

IL-6 is a soluble mediator with a pleiotropic effect, promptly and transiently produced in response to infections and tissue injuries; it contributes to host defence through the stimulation of acute phase proteins, haematopoiesis, and immune reactions (Tanaka et al., 2014). We reviewed eleven studies analyzing serum levels of IL-6 in correlation with cognition (Barbosa et al., 2018; Chakrabarty et al., 2019; M. H. Chen et al., 2021; Hamdani et al., 2015; Hoseth et al., 2016; K. L. Huang et al., 2022; King et al., 2019; Mora et al., 2019; Poletti et al., 2021; Rahmani et al., 2021; Wiener et al., 2017). Hamdani et al. (2015) investigated the presence of Toxoplasma gondii, an inducer of IL-6 expression, in euthymic BD patients and healthy controls. Patients and controls were evaluated for episodic verbal memory using the California Verbal Learning Test (CVLT) and for working memory and verbal ability using the Vocabulary subtest of the Wechsler Adult Intelligence Scale III (WAIS-III). A cognitive deterioration index was calculated based on subtests from the WAIS-III combining tests that are less sensitive to deterioration (e.g., Vocabulary) and tests known to be more sensitive to deterioration (e.g., Digit span). The seropositivity for Toxoplasma gondii and levels of IL-6 mRNA expression were significantly higher in BD patients than in controls. Levels of IL-6 mRNA expression were positively correlated with the cognitive deterioration index in BD patients. In Barbosa et al. (2018), 20 remitted patients with BD type I and 25 gender-matched controls were studied. Plasma levels of various inflammatory markers (i.e., sTNFR1, TNFα, IFN, IL2, IL4, IL6, IL10, IL17) were measured. In this study, patients with BD had elevated inflammatory markers when compared with controls, but IL-6 was the only cytokine directly associated with cognitive impairment assessed with the Brief Assessment of Cognition in Affective Disorders (BAC-A; total score). Similarly, M.H. Chen et al. (2023), who studied patients with first-episode schizophrenia, bipolar disorder, major depressive disorder, and healthy controls, reported that among the assessed cytokines (TNF-α, IL-6, IL-2), only CRP levels (P = .027) were negatively associated with WCST scores. Poletti et al. (2021) measured serum levels of a broad panel of inflammation markers and investigated their association with cognitive functioning in patients with a depressive episode in course of bipolar disorder. Patients performed a neuropsychological evaluation with the Brief Assessment of Cognition in Schizophrenia: a broad battery evaluating several domains of cognition (verbal memory, working memory, psychomotor speed and coordination, selective attention, semantic fluency, letter fluency and executive functions). Higher levels of IL-1β, IL-6 and chemokines were associated with the likelihood of having a poor cognitive performance. In another study, Rahmani et al. (2021) examined serum levels of CRP, IL-1β, IL-6, IL-10, IL-4 and TNF-α in adolescents with BD and healthy controls. Pro- to anti-inflammatory marker ratios were also examined. Neurocognitive flexibility was assessed via the intra/extradimensional shift (IED) task from the CANTAB battery. Within symptomatic BD adolescents, but not asymptomatic BD or HC adolescents, lower IL-6/IL-10 and lower CRP/IL-10 ratios were significantly associated with worse performance on the neurocognitive flexibility task. However, the other studies reviewed did not show significant correlation between IL-6 and cognition (Chakrabarty et al., 2019; M.H. Chen et al., 2020; Hoseth et al., 2016; M. H. Huang et al., 2022; King et al., 2019; Mora et al., 2019). Nevertheless, participants with BD had peripheral interleukin 6 levels that were positively correlated with functional impairment, suggesting that IL-6 may be involved in the neuroprogression of BD (Wiener et al., 2017).

Other ILs

IL1Ra is a protein secreted by various types of cells including immune cells. It binds non-productively to the cell surface interleukin-1 receptor preventing the pro-inflammatory effect of interleukin 1 alpha (IL1α) and interleukin 1 beta (IL1β) (Gabay et al., 2010). Notwithstanding findings that IL-1RA can block the cognition impairing effects of IL-1β, there is also evidence that IL-1RA may worsen cognition. Three studies analysed serum levels of IL-1RA in BD patients, showing a correlation with cognitive performance in two of them (Hope et al., 2015; Hoseth et al., 2016; Lotrich et al., 2014). In a cross-sectional correlational study measuring multiple cognitive domains, IL-1RA was found to be elevated in euthymic BD subjects compared to age-matched controls and, more importantly, was found to be associated with worse performances with regard to three cognitive domains: processing speed, executive function and cognitive control (Lotrich et al., 2014). Similar findings emerged from another study in which IL-1RA was negatively associated with both perceptual organization and verbal comprehension in BD patients (Hope et al., 2015). However, in another study conducted by Hoseth et al. (2016) there were no significant associations between IL-1RA, and verbal memory learning and recall tests. The IL-18, also known as interferon-gamma (IFNγ) inducing factor, is a pro-inflammatory cytokine that mainly acts on T cells and natural killer (NK) cells to induce production of IFNγ, which plays an important role in activating the macrophages.

IL-18 may provoke severe inflammatory responses, indicating its involvement in various inflammatory conditions, including chronic inflammation and autoimmune disorders (Ihim et al., 2022).

Neuroimaging abnormalities

In the neuroimaging studies identified by our search, the most consistently used neuroimaging technique was structural MRI (16 studies); five studies used Diffusion Tensor Imaging (DTI); four studies used resting-state functional magnetic resonance imaging (rs-fMRI). The results are summarized in table 2. The fronto-limbic structures were the region of interest most studied in association with inflammatory markers.

Table 2.

Summary of Neuroimaging Studies Investigating Inflammatory Markers in Individuals with Bipolar Disorder (BD)

STUDY	SUBJECTS DESCRIPTION	AGE	AGE	IMAGING	MISURES OF COGNITIVE PERFORMANCE	MAIN FINDINGS
Bai et al. (2020) Journal of Neuroinflammation	72 BD (45F/27M) 64 UD (45F/37M)	BD 39.5±12.3 UD 39±15.3	TG HDL LEP GHR INS ADI SG CHOL sIL-6R sIL-2R CRP sTNFR1	MRI		Patients with bipolar disorder (BD) have elevated levels of pro-inflammatory cytokines (sIL-6R, sTNF-R1) in comparison to individuals with unipolar depression (UD). These heightened cytokine levels are associated with changes in grey matter volume in specific brain regions and a positive correlation with a higher body mass index (BMI).
Benedetti et al. (2016) Journal of Affective Disorders	31 BD (21F/10M)	BD 48 ±13.09	TNF-α IFN-γ IL-5 IL-6 IL-7 IL-8 IL-10 MCP1 CXCL10 PTX3 G-CSF VCAM-1 ICAM-1 IL-1RA IL-2RA BDNF S100B SCF IFGBP-2 EGF PDGF-BB VEGF	DTI		Elevated cytokine levels in bipolar disorder (BD) were linked to reduced white matter integrity, as indicated by lower FA values. Specifically, TNF-α and IFN-γ had negative effects on oligodendrocytes and myelin, while IL-10 demonstrated a protective role.
Besga et al. (2017) Frontiers Aging Neuroscience	35 AD - Information not available 24 LOBD - Information not available 19 HC - Information not available	Above 50 y/o - Information not available	IL-1β IL-6 TNF-α PGE2 PGJ2 COX-1/2 NGF BDNF NO2 MDA	DWI		Late-Onset Bipolar Disorder (LOBD) displayed more white matter damage, inflammation, and oxidative stress compared to Alzheimer's Disease (AD), suggesting NGF and inflammation markers as potential diagnostic tools.
M.H. Chen et al (2019) Journal of Affective Disorders	23 BD (18F/5M) 23 UD (18F/5M)	BD 41.96 ±12.12 UD 43.35 ±13.62	sIL-6R sIL-2R CRP TNFR1	MRI		Bipolar disorder patients had smaller grey matter volumes in various brain regions than those with unipolar depression. This related to higher sIL-6R levels and an interaction with TNFR1 levels, suggesting a link between inflammation and brain differences.
P. Chen et al. (2020) Journal of Affective Disorders	22 BD (first episode) (18F/4M) 22 MMD (first episode) (18F/4M) 22 HC (18F/4M)	BD 28.1±12.0 MDD 28.7±13.3 HC 27.4±10.2	sIL-6R CRP TNFR1	MRI	Wisconsin Card Sorting test (WCST)	BD exhibited higher TNFR1 levels, a great deficit in executive function and a smaller grey matter (GM) volume in the middle frontal cortex (MFC) compared with MMD and HC. Correlation analysis revealed a positive association between the TNFR1 level and left MFC volume.
Chung et al. (2013) Neuropsychobiology	17 euthymic BDI (10F/7M)	BD 31.3 ± 7.0	hsCRP	MRI	Wisconsin Card Sorting test (WCST)	The left and right orbitofrontal cortex had a significantly negative relationship with serum hs-CRP levels. Analysis revealed a significant correlation between the left inferior orbitofrontal cortex and the indices of WCST performance.
Comai et al. (2022) European Neuropsychopharmacology	100 MDD (65F/35M) 66 BD (44F/22M)	MDD 49.48 ±10.99 BD 47.58± 8.55	IL-1β IL-1rα IL-2 IL-6 IL-8 IL-9 IL-10 IL-12 IL-13 IL-15 IL17 INF-γ TNF-α CCL2-CCL3 CCL4-CCL5 CCL11 CXC 10 FGF G-CSF GM-CSF PDGF-B VEGF TRP KYN	DTI		In bipolar disorder (BD), higher pro-inflammatory cytokine levels were associated with increased tryptophan degradation into kynurenine (Kyn) and reduced white matter (WM) microstructure. Conversely, in major depressive disorder (MDD), higher anti-inflammatory cytokine levels were linked to lower tryptophan degradation and improved WM microstructure, suggesting distinct mechanisms underlying WM disruption in BD and MDD.
Gong et al. (2022) Brain Imaging and Behavior	51 BD (15 BDI, 40BDII) (27F/24M) 69 HC (36F/33M)	BD 27.84 ±8.63 HC 31.48 ±11.62	IL-6 IL-1β TNF-α	fMRI		Patients with bipolar disorder (BD) displayed disrupted resting-state functional connectivity (rs-FC) between the amygdala and the medial prefrontal cortex (MFC) and temporal pole (TP). Additionally, BD patients exhibited increased levels of TNF-α, which negatively correlated with these abnormal rs-FC patterns. These findings suggest a potential role for TNF-α in influencing connectivity and emotional regulation in BD.
Hoseth et al. (2016) Acta Psychiatrica Scandinava	117 BD type I/II and not otherwise specified. (71F/46M) 109 SCZ (50F/59M) 236 HC (132F/104M)	BD 32 ±21 SCZ 30 ± 13 HC 35 ± 17	sTNF-R1 OPG IL-1RA I-L6 SCD40L CRP vWf	MRI	Wechsler Memory Scale (WMS-III): Logical Memory immediate (LM learning) and delayed recall (LM recall) California Verbal Learning Test (CVLT): learning and delayed free recall	Patients showed higher levels of sTNF-R1 and vWf compared to HC. There was a significant negative association between sTNF-R1 and verbal memory (LM learning and LM recall).
Lesh et al. (2018) Journal of Neuroinflammation	16 BD (first episode) (25F/75F) 69 SCZ (first episode) (12F/88M) 53 HC (36F/64M)	BD 21.4±3.4 SCZ 19.9±3.5 HC 19.5±3.3	IL-1β IL-6 IFN-γ IL-2 IL-10 IL-12	MRI	Wechsler abbreviated scale of Intelligence (WASI)	Elevated cytokine levels were observed in individuals with schizophrenia and bipolar disorder, with correlations between cytokines, symptom severity, and reductions in grey matter volume, indicating ongoing immune dysregulation in both conditions.
Lizano et al. (2019) The American Journal of Psychiatry	554 PP (SCZ 223, SCZ Affective 141, BD 190) (287F/267M) 522 FDR (364F/158M) 333 HC (184F/149M)	PP 35.5 ±12.5 FDR 40.8±15.6 HC 37±12.4	IL-1β IL-2 IL-4 IL-6 IL-8 IL-10 IL-12 IL-12p70, IFN-γ TNF-α CRP sFlt-1 VEGF TGF-β	MRI	Brief Assessment of Cognition in Schizophrenia (BACS)	Choroid plexus volume was significantly larger in individuals across the psychosis spectrum (schizophrenia, schizoaffective disorder, and psychotic bipolar I disorder) compared to healthy controls. Additionally, the study revealed a significant association between larger choroid plexus volume and elevated levels of the inflammatory cytokine IL-6 in individuals with psychosis, further implicating the neuroimmune axis in the structural and cognitive abnormalities observed in these disorders.
Lizano et al. (2021) Molecular Psychiatry	79 SCZ 61 Psychotic BD (PZ 60M/80 F) 60 HC (26M/34F)	PZ 34.3± 13.3 HC 37.6± 13.5	IL-1α /IL-1β IL-2 IL-4 IL-6 IL-8 IL-10 IL-23p40 IL-12p70 IFN-γ TNF- α / TNF-β CRP Flt-1 VEGF (C/D)	MRI	Brief Assessment of Cognition in Schizophrenia (BACS) Spatial Span subtest from Wechsler Memory Scale (WMS) Dot pattern expectancy task reaction time (DPX RT) Antisaccade error rate(ER)	Elevated levels of specific markers, including TNFα, IL6, VEGF and CRP, were associated with cognitive and neuroanatomical differences. Notably, higher IL6 and CRP levels were linked to poorer cognition in probands, while elevated CRP and VEGF levels were associated with increased antisaccade errors, but no such correlations were found in healthy controls.
Magioncalda et al. (2018) Brain, Behavior and Immunity	20 maniac BD (13F/7M) 20 depressed BD (11F/9M) 20 euthymic BD (12F/8M) 20 HC (9F/11M)	maniac BD 47.2±7.9 depressed BD 45.7±8.8 euthymic BD 48.2±10.1 HC 41.3± 14.1	IFN-γ IL-10, IL-4 IL-6 IL-17A IL-1β TNF-α CD4+ CD8+ CD28- CD45RA+ CD28- CD45RA+	DTI		Individuals experiencing mania in bipolar disorder exhibit widespread changes in white matter in the brain, particularly in regions called the BCC and SCRL. Significant alterations were observed in the immune system during manic episodes, with an increase in CD4+ T cells and a decrease in certain CD8+ T cell subtypes. These changes appear to be closely linked to the structural white matter abnormalities seen in mania.
Mohite et al. (2022) CNS Spectrums	21 BD (14F/7M) 22 HC (13F/9M)	BD 33.9±9.17 HC 33.91±9.99	IL-1β IL-2 IL-4 IL-6 IL-8 IL-10 TNF-α INF-γ	MRI	Montreal Cognitive Impairment (MoCA)	The study found negative correlations between medial orbitofrontal cortex volume and pro-inflammatory cytokine IFN-y levels, and positive correlations between posterior cingulate cortex volume and anti-inflammatory cytokine IL-10 levels in individuals with bipolar disorder (BD).
Papiol et al. (2008) Genes, Brain and Behavior	20 BD (10F/10M) 45 HC (24F/21M)	BD 43,3± 11.7 HC 29.4±9	IL-1β IL-1Ra -Genetic analysis	MRI		The study identified a genetic variant (allele*2) in the IL-1B gene that was linked to widespread grey matter deficits in bipolar disorder patients, with a pronounced effect in the left dorsolateral prefrontal cortex. These findings suggest a common genetic risk factor with schizophrenia and implies potential involvement of progressive neurodegenerative processes influenced by ischemic mechanisms and the hypothalamus- pituitary-adrenal axis.
Poletti et al. (2017) Brain, Behavior and Immunity	25 depressed BD (17F/8M) 21 HC (15F/6M)	BD 47.44± 12.99 HC 27.61± 10.17	IL-17 IL-22 CCR6+ CXCR3 CCR4 CCR10 CD161 IFN-γ Th1 Th2 Treg TNF-α IL-8 IL-10	fMRI		The study indicates that an imbalance between Th17 and T regulatory cells might be involved in bipolar disorder (BD) and related brain changes. It suggests that more Th17 cells could lead to better white matter (WM) structure, while increased T regulatory cells might be linked to weaker WM, and reduced brain responses to negative stimuli in BD patients.
Shonibare et al. (2020) Journal of Psychiatric Research	38 BD (23F/15M) 32 HC (19F/13M)	BD 17.5±1.2 HC 16.4±1.1	Genotype (IL-1β rs 16944)	MRI		In teenagers with bipolar disorder and healthy individuals, a specific genetic variation (IL-1β rs16944) was linked to larger brain regions. This suggests that this genetic difference might affect the brain's structure especially in frontal and occipital cortex relevant to BD. These results suggesting genetic variations in IL-1β may influence brain structure in adolescents with BD.
Strenn et al. (2021) Psychiatry Research	188 BD (114F/74M) 54 HC (32F/22M)	BD 38 (18-73) HC 40 (21-74)	IL-1β (rs1143623, rs1143627, rs1143634 and rs16944)	MRI		This study investigated genetic variations in the IL1B gene and their association with brain region volumes in controls and bipolar patients, revealing certain genetic variations (SNPs rs16944 and the C allele at rs1143627) linked to increased volumes in various brain areas, particularly in the left hemisphere putamen.
Tang et al. (2021) Psychological Medicine	42 BDII (23F/19M) 62 HC (36F/26M)	BDII 26.95 ±8.21 HC 31.48±11.62	IL-6 IL-8	fMRI		BD II patients showed decreased functional connectivity in brain regions related to the limbic and somatomotor networks compared to healthy controls, along with increased IL-6 and IL-8 levels, suggesting a potential link between inflammation and brain functional abnormalities in BD.
Tsai et al. (2019) Journal of Affective Disorders	32 BD -Information not available	BD 61.2±8.3	IL-1Ra IL-1β sIL-2R sIL-6R sTNF-R1	MRI		Older bipolar patients had elevated levels of peripheral inflammatory markers, including sIL-2R, sTNF-R1, and IL-1β, which were associated with reduced volumes of the bilateral hippocampi and total grey matter. Higher BMI and inflammation were linked to smaller left hippocampal volume, suggesting the impact of aging and obesity on neuroinflammation in BD.
Tsai et al. (2022) Acta Neuropsychiatrica	31 euthymic BDI (15F/16M) treatment 14 lithium 10 valproate 7 lithium and risperidone 3 valproate and quetiapine	BD 29.5±6.5	MCP-1 YKL-40 FKN sTNF-R1 IL-1β TGF-β1	MRI		Plasma levels of YKL-40 and sTNF-R1 were negatively associated with brain subarea volumes in bipolar disorder patients, particularly in frontal, temporal and parietal lobes, indicating inflammation's potential role in brain volume reduction in BD-I.
Tseng et al. (2021) Brain, Behavior, and Immunity	25 euthymic BD (16F/9M) on medication 43 HC (22F/21M)	BD 36.48± 12.69 HC 31.5±9.96	Hs-CRP	fMRI	Wisconsin Card-Sorting test (WCST) Continuous Performance test (CPT)	BD patients had higher hs-CRP levels than controls. This elevated inflammation level was associated with increased dorsal corticostriatal connectivity, specifically from the dorsal striatum to the vlPFC, in BD patients. These altered connectivity patterns were linked to cognitive deficits in attention and executive function observed in BD patients.
Tu et al. (2017) Journal of Affective Disorders	75 BD (35 BD I, 30 BD II) (48F/27M)	BD 42.67 ± 10.39	sIL-6R	MRI	Wisconsin Card-Sorting test (WCST)	Elevated serum sIL-6R levels in bipolar disorder patients were linked to thinning of the left middle temporal cortex and increased connectivity between the medial prefrontal cortex and limbic structures. This study provides insights into the association between inflammatory markers and structural and functional brain abnormalities in bipolar disorder.
Vai et al. (2022) Brain, Behavior & Immunity-Health	84 BD (64 depressed, 16 maniac, 4 euthymic) (29F/55M)	BD 47.01 ± 11.85 Depressed 46.39 ±11.52 Maniac 52.31 ±11.41Euthymic 37.25±12.61	IL-6	MRI		In BD patients, a higher level of peripheral IL-6 was associated with lower GM volumes in supragenual ACC and reduced rs-FC between this area and medial OFC. Additionally, higher IL-6 levels were correlated with cognitive biases associated with depressive episodes.
Wei et al. (2023) Psychological Medicine	396 SCZ 286 BD 376 MDD 711 PZ structural connectivity dataset (458F/253M) 559 PZ functional connectivity dataset (383F/176M) 485 HC: 313 HC structural connectivity dataset (183F/130M) 280 HC functional connectivity dataset (174F/106M)	PZ Structural dataset 27.36± 10.25 PZ functional dataset 26.18±9.54 HC structural dataset 28.6±10.23 HC functional dataset 27.73±8.3	IL-6 BR13 IL-1β BR28 TNF-α BR12	DTI	Wisconsin Card-Sorting test (WCST)	Executive deficits in individuals with major psychiatric disorders (MPD) were correlated with decreased fractional anisotropy (FA) in specific brain regions, increased regional homogeneity (ReHo) in other regions, and elevated levels of pro-inflammatory cytokines IL-1β and IL-6. Fronto-limbic-striatal regions may represent a potential predictor and a promising avenue for therapeutic interventions.

BD: Bipolar Disorder; LOBD: Late-onset Bipolar Disorder; UD: Unipolar Depression; AD: Alzheimer's Disease; PP: Psychosis Probands; FDR: First-degree Relatives; MRI: Magnetic Resonance Imaging; DTI: Diffusion Tensor Imaging; DWI: Diffusion Weight Imaging; FA: Fractional Anisotropy; rs-FC: Resting-State Functional Connectivity; BCC: Body of the Corpus Callosum; SCRL: Superior Coronal Radiata Left; WM: White Matter; LEP: Leptin; GHR: Ghrelin; INS: Insulin; ADI: Adiponectin; SG: Serum Glucose; TG: Triglyceride; CHOL: Cholesterol; IL-1β: Interleukin-1β; IL-1rα: Interleukin-1 receptor alpha; IL-2: Interleukin-2; IL-4: Interleukin-4; IL-6: Interleukin-6; IL-8: Interleukin-8; IL-9: Interleukin-9; IL-10: Interleukin-10; IL-12: Interleukin-12; IL-13: Interleukin-13; IL-15: Interleukin-15; IL-17: Interleukin-17; IL-1RA: Interleukin-1 Receptor Antagonist; IFN-γ: Interferon gamma; TNF-α: Tumor Necrosis Factor alpha; CCL2: C-C Motif Ligand 2; CCL3: C-C Motif Ligand 3; CCL4: C-C Motif Ligand 4; CCL5: C-C Motif Ligand 5; CCL11: C-C Motif Ligand 11; CXCL10: C-X-C Motif Chemokine 10; CCR6+: C-C Chemokine Receptor Type 6 Positive; CXCR3: C-X-C Chemokine Receptor Type 3; CCR4: C-C Chemokine Receptor Type 4; CCR10: C-C Chemokine Receptor Type 10; CD161: Cluster of Differentiation 161; Th1: T-Helper 1; Th2: T-Helper 2; Treg: T Regulatory Cells; PGE2: Prostaglandin E2; PGJ2: Prostaglandin J2; COX-1/2: Cyclooxygenase-1/2; NGF: Nerve Growth Factor; BDNF: Brain-Derived Neurotrophic Factor; NO2: Nitrite; MDA: Malondialdehyde; TRP: Tryptophan; KYN: Kynurenine; MCP1: Monocyte Chemotactic Protein 1; PTX3: Pentraxin-Related Protein 3; VCAM-1: Vascular Cell Adhesion Molecule 1; ICAM-1: Intercellular Adhesion Molecule 1; S100B: S100 Calcium Binding Protein B; SCF: Stem Cell Factor; IFGBP-2: Insulin-like Growth Factor-Binding Protein 2; EGF: Epidermal Growth Factor; PDGF-BB: Platelet-Derived Growth Factor-BB; VEGF: Vascular Endothelial Growth Factor; FGF: Fibroblast Growth Factor; GM-CSF: Granulocyte Macrophage Colony Stimulating Factor; PBMC: Peripheral Blood Mononuclear Cells.

Structural neuroimaging abnormalities

Several studies investigated associations between specific grey matter (GM) volume changes of the brain and levels of inflammatory markers. In one study on euthymic patients with BD, CRP were negatively correlated with a reduced volume of the orbitofrontal cortex in both hemispheres (Chung et al., 2013). The regional deficits in the orbitofrontal cortex were associated with poor WCST performance, suggesting an association between CPR levels and executive dysfunction in euthymic bipolar patients. M.H. Chen et al. (2019) explored the association between pro-inflammatory cytokines (i.e., soluble interleukin-6 receptor (sIL-6R), soluble interleukin-2 receptor (sIL-2R), C reactive protein (CRP), soluble tumour necrosis factor receptor type 1 (TNF-R1)) and differences in brain grey matter (GM) volumes between BD and depressive disorder. Voxel-based morphometry (VBM) was performed to investigate the grey matter volume differences. Compared to patients with depression, BD individuals had significantly higher levels of sTNF-R1 than patients with depression, whereas levels of sIL-2R, sIL-6R, and CRP did not differ between unipolar depression and bipolar disorder. Several brain regions, including the right orbitofrontal cortex, the right inferior frontal cortex, the left cingulate cortex, the left frontal pole, the left middle frontal cortex, the left Heschl's gyrus, and the bilateral lingual gyri, were significantly smaller in patients with bipolar disorder than in patients with unipolar depression. Orbitofrontal cortex, lingual gyrus, inferior frontal cortex, middle frontal cortex, and planum polar were negatively correlated with the sIL-6R levels, but they were not correlated with the sIL-2R and sTNF-R1 levels. In another study, the same author (M.H. Chen et al., 2020) investigated the relationship between the grey matter volume (GM), analysed with VBM, and pro-inflammatory cytokine receptors (soluble interleukin-6 receptor and sTNF-R1 in patients with first-episode bipolar disorder, patients with first-episode major depressive disorder, and controls. All patients underwent comprehensive assessments, including clinical assessment, executive function examination (WCST). Patients with bipolar disorder exhibited higher levels of sTNF-R1 and smaller GM volume in the middle frontal cortex compared with those with major depressive disorder and healthy controls. Positive associations were observed between the middle frontal cortex volume, executive function, and the TNF-R1 level. In a similar study conducted by Bai et al. (2020) BD patients had higher sIL-6R and sTNF-R1, and reduced GM volumes over 12 areas (i.e., right cerebellar lobule, left and right putamen, left and right superior frontal gyrus, left lingual gyrus, left precentral gyrus, right fusiform gyrus, left calcarine, right precuneus, left inferior temporal gyrus, left hippocampus). The results indicate that the pro-inflammatory cytokines may be related to the greater GM volume reduction in bipolar disorder compared with unipolar disorder. The structural GM volume reductions in patients with BD negatively correlated with sIL-6R and sTNF-R1 levels. Tu et al. (2017) enrolled 74 stable patients with BD-I or BD-II. Structural MRI, using an automated surface-based method (FreeSurfer) to measure cortical thickness, were performed in all participants. Higher sIL-6R levels were associated with a thinner cortex in the right middle temporal gyrus. Another study investigated the relationship between specific brain regions and the total GM volume and peripheral inflammatory marker in older patients with BD (aged ≥50 years). Magnetic resonance imaging segmentation methods were used for visualizing the brain's anatomical structures. Each brain region was divided by an individual's total intracranial volume to obtain the brain region's volume in percentage relative to the total intracranial volume (calculated by the summation of the GM, white matter (WM), and CSF volumes). The plasma levels of soluble tumour necrosis factor receptor-1 (sTNF-R1), soluble interleukin (IL)-2 receptor (sIL-2R), sIL-6R, IL-1β, and IL-1 receptor antagonist were measured. Regression analyses revealed that the right hippocampal volume was negatively associated with the levels of sIL-2R and sTNF-R1. The left hippocampal volume was negatively associated with the sIL-2R level, while the total GM volume had an inverse relationship with sTNF-R1 and IL-1β levels (Tsai et al., 2019). In a more recent study, the same author investigated the association between peripheral inflammatory markers and brain subregion volumes in BD euthymic patients with BD type I (aged 20-45). Higher chitinase-3-like protein 1 (YKL-40) and sTNF-R1 levels were both significantly associated with lower volumes of the left anterior cingulum, left frontal lobe, right superior temporal gyrus, and supramarginal gyrus (Tsai et al., 2022).

Vai et al. (2022) reported that higher level of peripheral IL-6 was associated with lower GM volumes in the supragenual anterior cingulate cortex in BD. The study, conducted by Mohite et al. (2022), enrolled euthymic patients with BD and healthy controls. High-resolution structural brain images were acquired, and measures of cortical thickness and GM volume were extracted. Plasma levels of interleukins (IL-1β, 2, 4, 6, 8, 10), TNF-α, and IFN-γ were assessed. No significant differences were found between BD patients and controls regarding any of the cytokine levels assessed. With respect to the volumes of the brain structures analysed, compared to controls, BD patients had significantly lower volumes in the cingulate, medial-orbitofrontal (MOF) and parahippocampal regions. A negative correlation was found between medial-orbito-frontal volumes and levels of IFN-γ, while posterior cingulate volumes were positively correlated with levels of the anti-inflammatory cytokine IL-10. These findings support an association between inflammatory cytokines and fronto-limbic GM abnormalities. Lizano et al. (2019) reported associations between larger volumes of choroid plexus and higher levels of IL-6 and IL-10 in patients with psychotic bipolar I disorder. More recently, the same research group hypothesized that a combination of microvascular and inflammatory markers would perform better than individual markers at predicting cognitive dysfunction and brain structural alterations in patients with psychosis. They examined the association of inflammatory marker patterns with brain structural abnormalities in patients with psychosis, including patients with psychotic bipolar disorder. Probands elevated inflammatory markers levels (notably, CRP, IFNγ, IL1β, IL8, IL10, TNFα, and VEGF) showed reduced cortical thickness in the right rostral middle frontal, bilateral medial orbital frontal and left inferior temporal regions compared to the healthy controls with low inflammatory markers levels (Lizano et al., 2021). In contrast to these studies, no correlations were observed by Lesh et al. (2018) between plasma levels of interleukin (IL)-1β, IL-2, IL-6, interferon (IFN)-γ and whole GM in bipolar disorder I patients with psychotic features, despite bipolar patients showing elevated inflammatory markers compared to controls. Moreover, in a previous study with non-euthymic patients, even though researchers found a negative association between sTNF-R1 and performance on verbal memory (Logical Memory immediate (LM learning) and delayed recall (LM recall) tests) in patients with bipolar spectrum disorders, no significant associations were found between sTNF-R1 or other inflammatory markers (e.g., IL-1RA; IL-6; high-sensitivity C-reactive protein) and hippocampal subfield volumes (Hoseth et al., 2016). The effects of individual differences and pro-inflammatory genetic polymorphisms on structural brain change provide us with new ways of understanding the processes leading to cognitive impairments in BD. One neuroimaging study has reported that a genetic variant in IL-1β (i.e., -511C/T polymorphism (rs16944) of the IL-1β gene) was associated with GM volume alterations in BD. Patients who were T-carriers of IL-1β gene polymorphism show a significant decrease in whole-brain GM and left dorsolateral prefrontal GM compared to non-carriers (Papiol et al., 2008). Other findings contrast the notion that increased inflammation leads to a reduction in brain volume, given that a haplotypic combination of polymorphisms in the IL-1β gene has been associated with increased GM volumes in key regions involved in mood regulation. Shonibare et al. (2020), who examined the effects of pro-inflammatory single nucleotide polymorphism IL-1β rs16944 on brain structure in adolescents with BD and healthy controls, found that T-carriers had larger lateral occipital volume and surface area compared to non-carriers, whereas the opposite was found in healthy controls. More recently, another study aimed to analyse the impact of IL-1 polymorphism on brain morphology found that genetic variants in IL-1β (i.e., T allele at rs16944 and the C allele at rs1143627) were associated with increased volumes of the putamen of the left hemisphere in patients with BD and controls (Strenn et al., 2021). We found only six diffusion tensor imaging (DTI) studies which analysed the association between peripheral inflammatory markers and measures of white matter microstructure. Besga et al. (2017) compared a group of Late Onset Bipolar Disorder (LOBD) patients, a group of patients with Alzheimer's disease (AD), and a group of healthy controls, to study significant differences in the white matter (WM) structure and their correlation with plasma biomarkers. Blood plasma samples were analysed to extract peripheral biomarkers of inflammation (i.e., IL-1β, IL-6, TNFα), oxidative stress, and neurotrophins. The authors found a widespread reduction in fractional anisotropy (FA) in LOBD relative to AD in a distributed fronto-temporal-occipital circuit. Moreover, they found a positive correlation between IL-1β and IL-6, and white matter alterations in LOBD, and a negative correlation in AD, suggesting that inflammatory activity in BD patients is higher than in AD patients. Benedetti et al. (2016) showed a negative correlation between TNF-α, IL-8, IFN-γ and IL-10 with lower fractional anisotropy (FA), higher mean diffusivity, and radial diffusivity (RD) within a large-scale network including corpus callosum, cingulum, superior and inferior longitudinal fasciculi, inferior fronto-occipital fasciculi, uncinate, forceps, corona radiata, thalamic radiation, internal capsule in patients with a major depressive episode in BD. Recently, Comai et al. (2022) investigated whether peripheral levels of immune/inflammatory markers were related to WM integrity in patients with depression and bipolar disorder. They found that IL-1β correlated with lower FA in the corpus callosum and the inferior fronto-occipital fasciculus in BD patients, suggesting that distinct mechanisms underly the pathophysiology of BD and major depression. However, transdiagnostic associations between brain dysconnectivity within fronto-limbic-striatal regions (decreased FA), inflammation (elevated IL-1β and IL-6 levels), and executive deficits across major psychiatric disorders, including schizophrenia, bipolar disorder, and major depressive disorder, were observed by Wei et al. (2023). The study by Vai et al. (2022) examined the association between IL-6 and brain structure/ function in euthymic patients with bipolar disorder (BD). They found that higher IL-6 levels were linked to reduced grey matter volumes in the supragenual anterior cingulate cortex (ACC) and decreased resting-state functional connectivity (rs-FC) between ACC and medial orbitofrontal cortex (OFC) in BD.

Some studies provide valuable insights into immune pathway dysregulation and their implications for structural abnormalities in bipolar disorder. Poletti et al. (2017) investigated the levels of circulating T cells in patients affected by bipolar depression and healthy controls. The percentage of circulating T helpers (i.e, Th17) correlated positively with higher FA in fiber tracts in the body of corpus callosum, inferior fronto-occipital fasciculus and other WM tracts bilaterally, including the splenium of the corpus callosum, superior and inferior longitudinal fasciculi, forceps, and anterior thalamic radiation, both in patients and healthy controls, while the frequency of circulating T regulatory cells correlated positively with higher radial and mean diffusivity in patients. Magioncalda et al. (2018) reported a significant correlation between reduced FA and increased RD in in the corpus callosum and corona radiata, with reduced frequencies of CD8+ terminal effector memory and CD8+ IFNγ+ T cells in mania. These DTI-immunological relationships were specific for the manic phase because they were not present in depression, euthymia, or healthy controls.

Functional neuroimaging (fMRI) abnormalities

Several studies also investigated associations between levels of inflammatory cytokines and brain functional connectivity using resting state functional MRI (rs-fMRI). Tu et al. (2017) found that higher sIL-6R levels were associated with increased functional connectivity (FC) between the medial prefrontal cortex (mPFC) and amygdala, pallidum, putamen, and insula, and decreased FC between the mPFC and subgenual anterior cingulate cortex, and frontal pole in stable patients with BD-I or BD-II. Similar findings have been reported by Vai et al. (2022) in patients during both acute depressive and manic episodes. Results showed that higher levels of peripheral IL-6 were associated with reduced rs-FC between the supragenual anterior cingulate cortex and medial orbito-frontal cortex in BD. Tseng et al., (2021) explored the association between CRP levels and resting state corticostriatal connectivity in euthymic BD. Compared with healthy controls, a higher connectivity between the right dorsal caudal putamen and the ventral lateral prefrontal cortex in the BD group was significantly correlated with a higher CRP level. The participants also completed the Wisconsin Card-Sorting Test (WCST) and the Continuous Performance Test (CPT). The authors found that compared to controls, stronger resting state corticostriatal connectivity in the BD group was correlated with lower CPT performance. This suggests that the effects of CRP on connectivity may interfere with cognitive function. In another study, P. Chen et al. (2020) acquired resting-state functional magnetic resonance imaging (rs-fMRI) data from patients with unmedicated BD II (current episode depressed) and healthy controls. Additionally, the serum levels of pro-inflammatory cytokines, including IL-6 and TNF-α, were detected. Patients exhibited decreased FC between the right posterior insula and the left postcentral gyrus, and increased FC between the left anterior insula and the bilateral insula (extended to the right putamen) when compared with the controls. Moreover, the patients showed higher IL-6 and TNF-α levels compared to controls. Furthermore, there was a negative correlation between IL-6 levels and the FC between the right posterior insula and the left postcentral gyrus. More recently, in a similar study with unmedicated patients with BD II (current episode depressed) and healthy controls, Tang et al. (2021) found that, compared with the controls, BD II patients showed decreased FC in the left orbitofrontal cortex (OFC) and the right precentral gyrus, and increased levels of IL-6 IL-8 levels. Abnormal FC in the right precentral gyrus was inversely correlated with IL- 8, whereas no significant correlation was found between FC in the left OFC and cytokines levels. In another recent study, Gong et al. (2022) used rs-fMRI to investigate the amygdala FC from two pairs of amygdala seed regions (the bilateral lateral and medial amygdala) in unmedicated BD patients and healthy controls. The levels of pro-inflammatory cytokines including interleukin IL-1β, IL-6, and TNF-α were measured in the serum. Patients with BD had elevated levels of TNF-α. They also demonstrated reduced functional connectivity from the right medial amygdala to the bilateral medial frontal cortex, and from the left medial amygdala to the left temporal pole and the right orbital inferior frontal gyrus, compared to healthy controls. Correlation analysis showed negative correlation between the TNF-α level and abnormal FC of the right medial amygdala-bilateral MFC, as well as a negative correlation between TNF-α levels and abnormal FC of the left medial amygdala-left TP in the BD group. These studies suggest that immune dysregulation might play a critical role in connectivity abnormalities in limbic and somatomotor networks implicated in BD.

Discussion

This review aimed to examine the literature exploring the relationship between markers of inflammation, cognitive functioning, and structural/ functional abnormalities (neuroimaging studies) among individuals diagnosed with bipolar disorder. Most of the studies reviewed concur on the notion that inflammatory state, measured by peripheral blood levels of CRP and cytokines, might be an important contributor of cognitive impairment observed in patients with BD. However, such studies have produced remarkably inconsistent results. There are many factors that might have contributed to these discrepancies. Firstly, these studies were performed on patients in various phases of the illness that may present a differential extent of cognitive impairment and immune system dysregulation. Indeed, research examining the issue of cognitive impairment in mood disorders has revealed that the degree and quality of cognitive impairment vary markedly between different states of the disease (Douglas et al., 2021). Abnormalities in emotional processing may be more closely related to the particular mood state in mood disorders, whereas cold cognition has been found to persist into remission or euthymia. Furthermore, different cytokine profiles appear to be associated with different mood states (i.e., differing profiles during periods of depression, mania, and euthymia) (Modabbernia et al., 2013). BD is associated with low-grade inflammation even during euthymic periods, with periods of increased inflammation associated with mood episodes (i.e., mania, depression). Therefore, longitudinal studies are necessary to monitor inflammatory markers within the same cohort of BD subjects to ascertain the relationship between these markers and the variations in cognitive measures across different phases of the illness. This substantial heterogeneity also suggests that inflammation could be a significant pathogenetic factor for only a subset of BD (inflammatory BD), it being pathophysiologically different from other subtypes of BD. Furthermore, several studies failed to control for potential confounding factors associated with the pro-inflammatory state the patients. Indeed, early life-stress, illness duration, type and duration of treatment and medical conditions (e.g., cardiovascular disorders, chronic inflammatory diseases) have been shown to significantly impact both the immune system and cognitive functioning (Lai et al., 2020; Miller et al., 2013; Van Exel et al., 2003). Furthermore, studies differ in sampling procedures (e.g., various types of biological material for the measurement of cytokine alterations including serum, plasma levels, mRNA) and methods used to assess cognitive performance (i.e., different types of cognitive tests) making it difficult to obtain comparable findings. At this point, it should be noted that many of the studies reviewed did not report variables such as the time of day that the sample was obtained. It is well known that functioning of the immune system exhibits peculiar circadian rhythms that are synchronized and coordinated by the suprachiasmatic nucleus (SCN). Consequently, the diurnal variation of inflammatory cytokines poses a risk of confounding if sampling is performed without regard to time of day (Zhou et al., 2010). Other limitations sometimes included small sample size and a lack of healthy controls. Therefore, disease-specific contribution of peripheral immune-inflammatory markers cannot be established.

Brain imaging analysis suggests that neuroinflammation in BD might be associated with structural and functional anomalies in various regions of the central nervous system. Structural imaging studies point to a significant role for cytokines and the immune response in brain volume integrity in in the frontal, temporal, and occipital cortices, and in WM abnormalities seen on diffusion tensor imaging in widespread areas, including fronto-limbic-striatal regions. Several studies using novel neuroimaging techniques, such as resting-state functional connectivity magnetic resonance imaging (rs-fcMRI) in BD, have been performed. These studies suggest that immune dysregulation might play a critical role in connectivity abnormalities in limbic and somatomotor networks implicated in BD. Disruption of these circuits may be responsible for the cognitive impairment observed in BD. Overall, findings from structural and functional analyses show heterogeneous results without a clear definition of structures involved, white-matter pathways, and functional connectivity abnormalities. Therefore, more research is needed, in particular to identify those altered structures and white-matter dysconnectivity associated with inflammation. The heterogeneity of the analytical methods used to explore the association between inflammation and brain imaging abnormalities, and the characteristics of the sample of each study, limit the conclusions. Moreover, it has been suggested that the progressive symptomatic and functional decline observed in some patients with BD may be associated with progressive neuroanatomical changes (e.g., cortical thickness, subcortical volumes). However, no studies have used a longitudinal design to assess the relationship between inflammatory/immune dysfunctions and structural/functional brain changes during the course of BD.

Several biologically plausible mechanisms have been identified to understand the impact of inflammatory signalling on cognitive function in BD. Cytokines can directly alter monoamine levels, microglia cells, and neuroplasticity, and activate the hypothalamic– pituitary–adrenal (HPA) axis (Rosenblat et al., 2018). Circulating cytokines traverse the blood-brain barrier (BBB) via active transport channels and through leaky regions (e.g., choroid plexus) of the BBB. In addition, recent findings in animal models suggest the presence of lymphatic vessels in the brain that could provide another direct entrance to CSN for cytokines (Louveau et al., 2015). Pro-inflammatory cytokines have been shown to alter monoamine levels. Particularly, IL-6 and TNF-α increase serotonin (5-HT) turnover, which may lead to depletion of 5-HT. IL-2 and interferon (IFNγ) are responsible for the selective activation of indolamine2,3-dioxygenase (IDO) that plays a major role in in serotoninergic neurotransmission. The sustained activation of IDO results in an increased catabolism of tryptophane, an essential amino acid precursor of serotonin (5HT), along the kynurenine pathway. This reaction reduces tryptophan availability for the synthesis of 5HT and, more importantly, increases the breakdown of tryptophan to neurotoxic tryptophan catabolites (TRYCATs) (i.e., kynurenine, kynurenic acid, and quinolinic acid). In addition, a substantial number of pre-clinical and clinical studies have shown that pro-inflammatory cytokines are responsible of the selective activation of guanosine triphosphate-cyclohydrolase-1 (GTP-CH1), which decreases the efficacy of tetrahydro biopterin (BH4), an essential co-factor of dopamine and serotonin biosynthesis which, in turn, decreases even further the availability of dopamine and serotonin in the brain (Rosenblat JD, Gregory JM, Flor-Henry S, 2018). Therefore, it is highly plausible that inflammation-induced decrease of 5HT and neurotoxic metabolites can directly impair cognition. Inflammation in BD has also been associated with overactivation of microglia. Microglia, under physiological conditions, play an important role in neuroplasticity, facilitating pruning in many brain regions (Frick et al., 2013). When microglia cells are stimulated by an inflammatory stimulus, they transform into and activate state leading to increased oxidative stress, further release of pro-inflammatory cytokines, neuro-apoptosis, and inappropriate synaptic pruning involving key regions sub-serving mood and cognition (Stertz et al., 2013). Another key mechanism by which inflammation may impair cognition in BD is the hypothalamic-pituitary-adrenal dysfunction (HPA) axis dysregulation. Elevated levels of pro-inflammatory cytokines (i.e., IFN, TNF-α, IL-6) activate the hypothalamic–pituitary–adrenal (HPA) axis, thereby increasing systemic cortisol levels (Beishuizen & Thijs, 2003). Finally, hypercortisolaemia can potently impact cognition through increasing hepatic tryptophan2,3-dioxygenase (TDO) activity, thereby increasing the breakdown of tryptophan to TRYCATs (Maes et al., 2011). Further, elevated levels of cytokines also decrease glucocorticoid receptors (GR) synthesis and sensitivity in the hypothalamus and pituitary, thus preventing the negative feedback loop which normally downregulates cortisol production.

Conclusion

Despite the contradictory results, these studies underly the importance of the role that neuroinflammation may play in the pathogenesis of cognitive impairment in BD. Particularly, robust evidence exists concerning an association between cognitive impairment and CRP, IL-1RA, IL-6, and TNF-α with its receptors, whereas there is no convincing evidence for the involvement of other neuroinflammatory biomarkers. Executive functions, attention, and memory are frequently found to be impaired in BD patients concomitantly with the presence of a pro-inflammatory state. Emerging evidence indicates that systemic inflammation might also play an important role in the deterioration of brain structures critical to cognitive functions in patients with BD. The identification of inflammatory markers underlying cognitive deficit and brain abnormalities in BD patients will have important implications for the development of specific therapeutic interventions targeting chronic inflammation.