Table 1.

Crosstalk between the GM and human brain function.

Subjects	Intervention	Measures	Results				References
			Neuroimaging results	GM results	Correlation	Other
Healthy women	FMPP for 4w (FMPP group, n = 12; non-FMPP group, n = 12; no intervention, n = 13)	Task-based fMRI and rs-fMRI; GM (fecal samples)	Emotional attention task based-fMRI: sensory brain network connection strength and decreases in insular and somatosensory cortical BOLD activity↓ in the FMPP group rs-fMRI: PAG was negatively correlated with sensory/affective regions and positively correlated with cortical regulatory regions (medial and dorsolateral prefrontal cortices) in FMPP group	No significant changes in fecal microbiota composition	Four-week intake of an FMPP by healthy women affected activity in brain regions that control the central processing of emotion and sensation	/	(25)
Obese and non-obese subjects	No intervention (20 obese and 19 non-obese subjects)	MRI; DTI; FLAIR; R2*; GM (fecal samples); cognitive tests	See the correlation column	16S bacterial gene pyrosequencing: fecal sample bacterial biodiversity↓ in obese men	Fecal microbiota diversity was negatively correlated with R2* signals in the hypothalamus, hippocampus, and caudate nucleus. The abundance of Actinobacteria was positively associated with FA in the amygdala and thalamus but negatively correlated with the R2* signal in the hypothalamus	The relative abundance of the Actinobacteria Phylum was positively associated with cognitive tests related to speed, attention, and cognitive flexibility	(26)
Elderly outpatients with and without cirrhosis	No intervention Group type 1: 39 cirrhotic and 37 non-cirrhotic patients; Group type 2: unimpaired cognition (n = 23), amnestic-type (n = 25), and amnestic/non-amnestic type (n = 28).	Multi-modal MRI (fMRI go/no-go task, volumetry, and MRS); inflammatory cytokines; GM (fecal samples); neuropsychological tests	No significant fMRI differences in brain volumes between cirrhotic and non-cirrhotic subjects. Amnestic/non-amnestic type: activation in the central opercular cortex, post-central gyrus, and superior parietal lobule during inhibition↑. Amnestic-type type: white matter, gray matter, and total brain volumes↓, hippocampal and left thalamic volumes↓ Cirrhotic subjects: mi/Cr and NAA/Cr ratios↓ and Glx/Cr ratio↑. Amnestic/non-amnestic type: mi/Cr and Glx/Cr ratios↓	Cirrhotic subjects: Lactobacillales↑ and Synergisticeae, and Peptococcaceae↓. Cognitively impaired groups: decreased Subdoligranulum, Oscillibacte, and Porphyromonadaceae and Prevotellaceae↓ and Bacteroides↑ Unimpaired group: increased Fecalibacterium and Butyricicoccus	Regardless of the presence of cirrhosis, beneficial taxa (Lactobacillales, Ruminococcaceae, and Lachnospiraceae) were positively linked with cognition while pathogenic taxa (Enterobacteriaceae) were negatively linked with cognition	Serum levels of IL-6/endotoxin↑	(27)
Cirrhosis patients with/without prior HE	No intervention (Cirrhotic without prior HE, n = 62; cirrhotic with prior HE, n = 85; controls, n = 40)	MRS, DTI, Systemic inflammatory assessment, GM (fecal samples); cognitive testing	MRS: Cirrhotic patients with HE: Glx↑↑ mi↓↓; Cho↓↓; Cirrhotics: Glx↑ mi↓; DTI: Cirrhotics with HE: spherical isotropy↑, FA↓	Cirrhotic patients with HE: Staphylococcaceae↑, Enterococcaceae↑, Porphyromonadaceae↑, and Lactobacillaceae↑; autochthonous bacterial families (Lachospiraceae↓, Ruminococcaeae↓, and Clostridiales XIV↓	Autochthonous taxa were negatively correlated with hyperammonemia-associated astrocytic MRS changes while Enterobacteriaceae were positively correlated with hyperammonemia-associated astrocytic MRS changes (high Glx levels and low mi and Cho levels). Porphyromonadaceae were correlated with neuronal changes on DTI without being linked to ammonia	Cirrhotic patients with prior HE had significantly more advanced cirrhosis, more severe and higher levels of inflammatory markers and cognitive impairments compared to cirrhotic patients without HE	(28)
IBS patients with anxiety and depression	Probiotic: Bifidobacterium longum NCC3001 for 6 w (BL, n = 22; placebo, n = 22)	Task-based fMRI (fearful face backward masking paradigm); fecal microbiota; urine metabolome profiles; serum inflammatory markers; neurotransmitter and neurotrophin levels	BL reduced responses to negative emotional stimuli in multiple brain areas including the amygdala and fronto-limbic regions	No major changes in fecal microbiota composition after intervention	Reduced amygdala activity was correlated with decreases in depression scores	Depression scores↓; IBS symptoms improved; urine levels of methylamines and aromatic amino acid metabolites↓	(29)
Healthy women (Bacteroides-high group vs. Prevotella-high group)	No intervention	Emotion-induced task-based fMRI; structural MRI (DTI, T1)	Prevotella group: right hippocampal activity↓ when viewing negative valence images. Bacteroides-high group: white matter connectivity↓, cerebellum, frontal region, and hippocampal volumes↑, and nucleus accumbal volume↓	Subjects were divided into Bacteroides-high group and Prevotella-high group based on fecal microbiota analysis	/		(30)
Obese and non-obese subjects	Diet counseling (18 obese and 17 non-obese subjects)	MR relaxometry R2*; GM (fecal samples); Neuropsychological tests; plasma β-amyloid (1–24, 31–48) levels	MR R2* relaxometry increased mainly in the pallidum, putamen, thalamus, and hippocampus in both groups over a 2-year period	A variety of gut microbiome changes in RA over a 2-year period	Shifts in Gemmatimonadetes, Bacteroidetes, Proteobacteria, Caldiserica, Candidatus, Saccharibacteria, Tenericutes, Thermodesulfobacteria, and Chlorobi RA were associated with increased percentages of R2* in the striatum, superficial amygdala, and hippocampus. Shifts in the phyla Fibrobacteres, Synergistetes, and Tenericutes RA were reciprocally associated with right hippocampal R2*.	Circulating β-amyloid Ab42 levels were positively associated with changes in visuospatial constructional ability and immediate memory but negatively associated with increases in R2*	(49)
Healthy volunteers	Probiotic (Ecologic825, nine bacterial strains) for 4w (probiotic, n = 15; placebo, n = 15; no intervention, n = 15)	Task-based fMRI (emotional decision-making and emotional recognition memory); DTI; neuropsychological tests	Altered brain activation in the cingulum, precuneus, inferior parietal lobule, thalamus, and parahippocampal gyrus in the ED task and cerebellar activity in the ER task	No major changes in general fecal microbial diversity or evenness; Bacteroides sp.↑, Alistipe sp.↑, and the nicotinate and nicotinamide metabolic pathway of fecal microbiota↓	Probiotic ingestion improved emotional attention and memory performance, which was accompanied by changes in activity in corresponding brain regions	Self-reported behavioral measures of positive affect, cognitive reactivity, and memory performance improved	(50)
Healthy volunteers	Probiotic (Ecologic825, nine bacterial strains) for 4w (probiotic, n = 15; placebo, n = 15; no intervention, n = 15)	Rs-fMRI; diffusion MRI	FC in MFGN (in frontal pole and frontal medial cortex) and in DMN (in frontal lobe)↓, FC in SN (in cingulate gyrus and precuneus cortex)↑. No significant changes in structural connectivity (FA/MD)	Same as above [Bagga et al. (50)]	Probiotic intervention was -associated with -changes in FC but not structural connectivity		(51)
High-risk (HR) and ultra-high risk (UHR) subjects for schizophrenia	No intervention (high-risk group, n = 81; ultra-high risk group, n = 19; healthy controls, n = 69)	MRS; GM (fecal samples).	Ultra-high risk group: Cho levels in anterior cingulate cortex↑	Ultra-high risk group: at order level; Clostridiales, Lactobacillales, and Bacteroidales↑; at general level, Prevotella and Lactobacillus↑; synthesis of acetyl-CoA (belonging to SCFAs)↑	Alterations in MRS and GM function (synthesis of SCFAs) support the hypothesis that membrane dysfunction exists in schizophrenia	/	(52)

Summary of multimodal MR imaging studies investigating the relationship between the adult GM and brain function.