The gut microbiota and mental health in adults

A growing body of evidence point toward the bidirectional gut microbiota-brain axis playing a role in mental health. Most of this research is conducted on animals why we in this review summarize and comment upon recent studies evaluating the gut microbiome in mental health in humans. Further support for the relevance of the bidirectional gut microbiota-brain communication in mood disorders has been presented, such as the effect of probiotics on brain connectivity and mental health outcomes and pregnancy related stress on gut microbiota in the newborn child. However, the heterogeneity between studies precludes conclusions regarding differences in microbiota composition in mental disease and health and many of the studies are limited by a cross-sectional design, small sample sizes and multiple comparisons. Thus, well-designed longitudinal studies with larger sample size, accounting for confounders are needed.


Introduction
Gut microbiota may influence brain function through neural, endocrine, and immune pathways [1]. For example, the gut microbiota may impair the integrity of the intestinal barrier. The resulting release of cytokines in turn may signal to the brain through vagal activation or signaling across the blood brain barrier. In addition, substances produced by the gut microbiota may be absorbed reaching the brain by the blood stream. The brain, in turn, may influence the gut microbiota through neuronal and endocrine pathways as well as through health behaviors.
The amount of investigations into the gut-microbiota-brain axis has escalated recently as the cost for microbiota analyses has decreased. Much of the research data available on the gut microbiota-brain axis communication and mental health sequela are derived from animal research [2]. In the present review we summarize the research advances in the field based on human research published during the last two years.
We searched for original articles and meta-analyses published 2018 and 2019 registered in PubMed until November 8th 2019. Search words included microbiota combined with brain, imaging, mental, anxiety, depression, stress disorder and bipolar disorder. Studies on multiple sclerosis, Parkinson's disease, Alzheimer's disease, epilepsy, autism spectrum disorder, case reports and studies only investigating children were not included. Only English-written manuscripts were included. In total, 43 original articles and two metaanalyses were found and included in this review, and the number of review articles was more than twice as many. The studies fell within four overarching categories: 1) brain imaging studies 2) depression 3) bipolar disorder and 4) anxiety/stress with some studies fitting into more than one category.

Brain imaging
Two randomized controlled trials investigated the effect of four-week probiotic treatment on brain activity in healthy volunteers. Participants receiving Bifidobacterium longum 1714TM differed from participants receiving placebo in the neural oscillations in, for example, frontal and cingulate cortex during both resting-state and after social stress and the changes observed during resting state were associated with higher self-reported vitality and reduced mental fatigue [3 ]. The other trial investigated a multispecies probiotic formulation. Participants receiving probiotics reported higher positive affect (mood state) and showed differences in the BOLD (blood-oxygen level-dependent) signal pattern in response to emotional decision-making and emotional recognition memory tasks compared to participants receiving placebo [4].
Three cross-sectional studies on brain connectivity were found. Higher resting state insular functional connectivity was associated with a higher fecal bacterial microbiota diversity, as well as with lower abundance of Bacteroides and higher abundance of Prevotella [5]. High levels of fecal indole metabolites were associated with functional and anatomical connectivity of areas associated with reward and anxiety [6]. Bacteroides, Parabacteroidetes and Escherichia species were identified as the source of GABA and an association between resting state functional connectivity in a neural circuitry considered important in depression and a lower abundance of Bacteroides in patient with major depression were found [7].
Above studies support brain imaging as a tool for investigating the link between gut microbiome and brain function with tentative support for an effect of probiotics on brain activity.

Depression
By March 2018 six studies investigating gut microbiota in relation to depression were available [8]. Since then an additional 15 studies have been published and are reviewed below. The key characteristics of these studies are summarized in Table 1.

Intervention studies Probiotic trials
Two meta-analyses evaluated the effect of probiotics in subjects with depressive symptoms and depression. The first meta-analysis included 10 clinical trials with a total of 1349 patients and found a significant mood improvement in individuals with pre-existing depressive mood symptoms after an eight-week trial of probiotics compared to placebo [9]. The second meta-analysis included three studies and 229 patients and found an overall positive effect of probiotics on depressive symptoms in clinically depressed patients when used as a supplement to antidepressant treatment compared to antidepressant treatment alone in two of the studies [10]. One of the included studies also reported a significant decrease in depressive symptoms in patients with mild to moderate major depressive disorder (MDD) receiving probiotics (Lactobacillus helveticus and B. longum) compared to patients receiving prebiotics [11].
Two randomized controlled clinical trials that were not included in the meta-analyses further evaluated the effects of probiotics on mood in patients with depression. Chahwan et al. did not find an effect of a multistrain probiotics on depressive scores but patients receiving the probiotics had a greater reduction in cognitive reactivity toward sad mood compared to patients receiving placebo [12]. Further, Rudzki et al. found that supplementing selective serotonin reuptake inhibitors (SSRI) antidepressant treatment with probiotic Lactobacillus plantarum 299v improved cognitive performance compared to SSRI with placebo in depressed patients [13 ]. Similarly, reduction in depressive scores, as well as 70% treatment response and 35% remission rate was noted in a group receiving supplemental probiotic (CBM588) to antidepressants compared with the group on antidepressant alone in an open-label trial in patients with treatment-resistant MDD [14].
Three trials evaluated the effect of probiotics on depressive symptoms in study populations other than patients with depression. Multispecies probiotics was found to improve impulsivity and decision-making compared to placebo, but not pain, quality of life, anxiety or depressive symptoms in a double-blind, placebocontrolled, randomized study in patients diagnosed with fibromyalgia [15]. Relatedly, a small 12-week openlabeled trial with Bifidobacterium infantis M-63 placebo found improved mental well-being compared to placebo among patients with IBS. In the same study, lower beta diversity was associated with distress and depression and lower Lachnospiraceae abundance was associated with a higher depression score while patients with anxiety were characterized by elevated Bacteroidaceae [16]. A nonrandomized open label study found that B. infantis M-63 given for three months improved mental well-being but not depression in individuals with IBS [17].
Taken together, the presented studies support the use of supplemental probiotics to relieve depressive symptoms, or at least improve related cognitive functions.

Fecal transplants
Two studies investigating the effect of fecal microbiota transplantation on depression scores in patients with IBS or other functional gastrointestinal disorders found fecal microbiota transplantation to alleviate depression and anxiety [18,19].

Case control studies
Three cross-sectional case control studies compared fecal microbiome in patients with MDD and healthy controls but results are discrepant. For example, a Taiwanese study reported that patients with MDD had an overrepresentation of phylum Actinobacteria and Firmicutes and the genus Bifidobacterium and Blautia compared to healthy controls [20]. In contrast, Huang et al. reported lower diversity in MDD, with lower abundance of Firmicutes compared to healthy controls [21]. A small metaproteomic study with age, sex and BMI matched controls found a different bacterial protein signature in MDD compared to controls, with differences in proteins belonging to glucose and amino acid metabolism. Considering the merits of this profile, the authors concluded that Firmicutes, Actinobacteria and Lachnospiraceae were more abundant in MDD, whereas Bacteroidetes and Proteobacteria were less abundant, in MDD compared to controls [22].
Two studies investigated gut microbiota products in relation to depression. Lower fecal levels of the bacterial metabolites short chain fatty acids (SCFA) acetate and propionic acid, but higher isocaproic acid concentrations were found in depressed compared to non-depressed Polish women [23 ]. In support of the theory of inflammatory changes by means of gut permeability, an American study found that individuals with more hostile marital interactions had higher bacterial endotoxin blood www.sciencedirect.com The gut microbiota and mental health in adults Jä rbrink-Sehgal and Andreasson 105 Individuals with more hostile marital interactions and history of mood disorder were observed to have a higher LBP/sCD14 ratios than those with less hostile interactions and without mood disorders.
Current Opinion in Neurobiology 2020, 62:102-114 www.sciencedirect.com The gut microbiota and mental health in adults Jä rbrink-Sehgal and Andreasson 107 Vegetarian IBD patients had lower mental SF-36 scores and higher level of post-traumatic stress symptoms. Significant differences in mucosaassociated gut microbiota composition was noted between vegetarian and gluten free IBD patients compared to IBD patients without diet restrictions.
www.sciencedirect.com levels compared to those less hostile, in particular those with a history of mood disorder [24 ].

Cohort studies
In the Flemish Gut Flora Project including 1054 subjects, fecal Dialister and Coprococcus spp. were depleted in depression, even after correcting for any confounding effects of antidepressants [25]. Heym et al. studied 40 participants from the general population in UK and found that the fecal abundance of Lactobacillus spp. was directly related to positive self-judgment but only indirectly to cognitive depression and lower affective empathy [26]. Two studies have investigated mucosa associated microbiota: In a Swiss cohort study higher depressive, anxiety and stress symptoms were associated with lower microbiota diversity in 171 patients with inflammatory bowel disorder (IBD) in remission [27].
In a Swedish population-based colonoscopy study evaluating 375 participants randomly selected from the general adult population, decreased diversity was associated with poorer self-rated health, but was not associated with anxiety or depression. However, several taxonomic groups were correlated to depression and self-rated health, including Coprococcus [28].
Two studies highlight the importance of diet for both gut microbiota composition and mental health. In adults without mood disorders, higher anxiety and lower fecal Bifidobacterium was associated in women and higher depression scores and lower Lactobacillus was associated in men. Importantly, an inverse association was noted between the total fruit and dairy ingestion and depression and stress scores, respectively [29]. In a study on IBD patients, an association between vegetarian diet and gluten free diet and higher levels of anxiety and depression symptoms was found, while the micobiota composition in patients following a vegetarian and gluten free diet differed significantly from that of omnivores, although the direct association between gut microbiota and mental health was not investigated [30 ].
In a small prospective observational study, Liskiewicz et al. found an increased fecal bacteria alpha biodiversity but no difference in abundance after a six-week inpatient treatment with SSRIs [31]. All patients achieved clinical improvement but the association between change in microbiome composition and the clinical course of depression was not investigated.
Overall the presented observational studies support an altered gut microbiome in subjects with depression and depressive symptoms. The specific gut microbiome composition linked to depression however varies among studies likely due to differences in study populations, methodology and presence of confounding factors.
108 Brain, gut, and immune system interactions

Bipolar disorder
Clinically, BD is characterized by alternating recurrent manic, depressive or mixed mood episodes throughout the course of disease [32]. In 2017, two studies assessing the gut microbiome in bipolar disorder (BD) had been published [33], while seven more has been published to date and are reviewed below, and their key characteristics are summarized in Table 2. The reviewed studies share similar methodology in terms of their cross-sectional case control design and fecal microbiome analysis but differ in the comparative groups used, ranging from healthy controls, first degree relatives and patients with MDD, degree of depressive symptoms in BD cases as well as microbiome analysis methodology.
Some studies have found BD to be associated with a decreased bacterial diversity when compared to healthy controls [34,35 ,36] while other studies found no difference [36,37]. The results are also inconsistent in terms of predominant phyla in BD [34,36,37,38 ,39]. BD patients with depressive symptoms have been found to have a less bacterial diversity compared to BD patients with euthymia [35 ]. Further, findings of higher levels of fecal Bacteroides, Clostridium, Bifidobacterium, Oscillibacter and Streptococcus when comparing individuals with BD with depressive symptoms, MDD and healthy controls [36] suggest depression to be a distinct factor in BD. In similar vein, depressed BD patients showed significantly higher abundances of Enterobacteriaceae, while already healthier BD patients showed higher abundances of Clostridiaceae and Roseburia [37]. Differences among types of BD, in terms of degree of mania, were reported by Rong et al.; BD-II, which is characterized by milder form of hypomania than BD-I, had higher abundance of Collinsella relative to subjects with BD-I [36]. Microbiome differences between patients with BD and depressive symptoms versus patients with MDD, have also been reported, namely an increased abundance of Fusobacteriaceae, Escherichia blattae DSM 4481 and Klebsiella oxytoca, but decreased B. longum subsp. infantis ATCC 15697 = JCM 1222 [36]. Importantly, treated patients were found to have a higher abundance of Klebsiella and Veillonella compared to untreated patients. Further, illness duration was associated with lower richness [36]. Other interesting findings within the BD field include reported associations between Lactobacillus counts and sleep and Bifidobacterium counts with lower cortisol levels [40].
Only one trial on probiotics in BD had been published. An open trial investigating the effect of a multistrain probiotic in euthymic individuals with BD found improvement in attention, psychomotor processing speed and executive function [41].
In brief, studies support an overall altered gut microbiome in BD but the specific microbiome composition and role of probiotics in BD remains unclear and underexplored.

Anxiety and stress
Eight studies were found that evaluated gut microbiota and stress disorders, anxiety or adverse life events.
Two clinical trials investigating the effect of probiotics where found. In healthy young adults, a randomized controlled trial with a 4-week trial of long-term use of Lactobacillus gasseri CP2305 (CP2305) found improvement in mental state, sleep quality, and gut microbiota under stressful conditions by attenuating the stress-induced decline of Bifidobacterium spp. and the stress-induced elevation of Streptococcus spp. [42]. Another randomized controlled trial evaluating the effect of multispecies probiotics on anxiety in healthy college students found improvement in panic anxiety, neurophysiological anxiety, negative affect, worry, and improved negative mood regulation in the probiotic group compared to placebo group [43].
Case control studies have reported that patients with generalized anxiety disorder have lower fecal bacterial The gut microbiota and mental health in adults Jä rbrink-Sehgal and Andreasson 111   The sources of heterogeneity between the reviewed studies [50]. alpha diversity, less operational taxonomic units and reduced Firmicutes and Tenericutes abundances [44] and decreased short chain fatty acid-producing bacteria and overgrowth of Escherichia-Shigella, Fusobacterium and Ruminococcus gnavus [45] compared to healthy controls. In addition, higher anxiety severity was associated with lower abundances of Eubacterium coprostanoligenes group, Ruminococcaceae UCG-014, and Prevotella 9, and higher abundances of Bacteroides and Escherichia-Shigella [44].
Veterans with combat-related post-traumatic stress disorder (PTSD) and cirrhosis had lower fecal microbial diversity, higher Enterococcus and Escherichia/Shigella and lower Lachnospiraceaeae and Ruminococcaceae compared to veterans with cirrhosis but without PTSD, and cognition was differently linked to gut microbiota in PTSD [46].
Three studies had been conducted on pregnant women. Pregnant women that reported exposure to two or more adverse childhood experiences had greater abundance of fecal Prevotella than pregnant women with none or only one adverse childhood experience [47]. One study investigated the hypothesis that prenatal psychosocial stress may affects child outcomes via the mother's intestinal microbiota and found an association between maternal pregnancy general anxiety and fecal microbiota composition but no associations with other stress measures included in the study [48]. Highlights among the publications is the report of an association between greater pregnancy-related anxiety and less diverse meconium microbiota community with lower abundance of the Enterococcaceae in the offspring [49 ], supporting a causal influence of mental health on gut microbiota composition.
In summary, the reviewed studies report improved anxiety and stress symptoms with use of probiotics in healthy subjects and altered gut microbiome in patients with generalized anxiety disorder compared to healthy controls.

Conclusion
The human research into the gut microbiota-brain axis and mental health has accelerated during the last two years. Overall the reviewed studies provide support for the relevance of the bidirectional gut microbiota-brain communication in mood disorders in humans, such as the effect of probiotics on brain connectivity and mental health outcomes and pregnancy related stress on gut microbiota in the newborn child. However, the research is still in its early stage and studies differ in terms of study design, study populations, microbiota analyses and outcomes ( Figure 1). This heterogeneity precludes conclusions regarding causality and differences in microbiota composition in mental disease and health. Importantly, this review further highlights the importance of confounding factors ( Figure 2) and limitations of cross-sectional designs, small sample sizes and multiple comparisons.

Future directions
In the quest of establishing causality and clarity in gut microbiome composition and mental health status, studies with larger sample sizes, longitudinal study designs with longer follow-up periods, consistent methodology ( Figure 1) and control for confounders ( Figure 2) are needed. In particular, studies in BD need to account for specific factors such as type of BD, presence of depression at time of study, illness duration and treatment. Additional studies with pathway oriented approaches are needed to help elucidate the mechanistics in the gut-brain communication, such as investigations of changes in brain connectivity in response to interventions or of markers of microbiota function, for example, bacterial metabolites in feces and blood in relation to mental health status and as outcomes in interventions. Hence, we propose adding brain imaging to future studies when relevant and feasible, to increase the understanding of the gut microbiota-brain pathways and potentially provide valuable information on 'responders' versus 'non-responders' in intervention studies. Only a few studies have investigated mucosa associated microbiota in relation to mental health status. Although considerably more difficult to obtain than fecal samples, mucosal microbiota likely gives a better representation of the 112 Brain, gut, and immune system interactions  • Age [28] • Sex [29,44] • Obesity status [6,28] • Personality [50] • Pharmacological treatment [23,36] • Illness duration [36] • Disease status [36] • Comorbidities [23] • Diet [29,30] • Smoking [5,44] • Alcohol intake [44] • Sleep [40] •Clinical variables

•Health behaviors Current Opinion in Neurobiology
Confounding factors identified in the reviewed studies. organisms in premium position to influence their host and a sigmoidoscopy may be considered when a full colonoscopy would not be feasible. Importantly, the reviewed gut microbiome studies investigated the bacterial microbiota, leaving a gap in the knowledge of unicellular and multicellular eukaryotic organisms as well as archaea to be filled in the future. Of special note, the quality of the process of collecting the samples is as important as that of the microbiota analysis for the quality of the result.

Conflict of interest statement
Nothing declared.