Neuroimaging Model Distinguishes Functional Constipation Subtypes with 89% Accuracy

The Neurological Architecture of Chronic Constipation

Functional constipation remains a prevalent gastrointestinal challenge in clinical practice, particularly among aging populations where the pooled prevalence reaches 8.5% and the odds ratio for patients aged 70 and older stands at 3.38 [1, 2]. While clinicians often manage the condition as a localized bowel dysfunction, it is frequently complicated by psychiatric comorbidities; specifically, anxiety and depression are associated with significantly higher prevalence (OR 3.16 and 2.74, respectively) and increased symptom severity [1, 3]. These comorbidities often render standard pharmacological interventions, such as 5-HT4 receptor agonists, less effective, prompting a shift toward therapies that address the microbiota-gut-brain axis (the bidirectional communication network between the enteric nervous system and the central nervous system) [4, 3, 5]. Despite the known interplay between emotional processing and gut motility, identifying which patients possess a distinct neuropsychiatric phenotype (a set of observable characteristics resulting from the interaction of neurological and psychiatric factors) remains difficult [6, 5]. A recent study of 187 patients utilized machine learning and structural magnetic resonance imaging to classify patients with and without comorbid anxiety and depression with 89.23% accuracy, identifying key structural features in the temporal pole, amygdala, and insula [7].

Stratifying the Gut-Brain Phenotype

To investigate the neurological underpinnings of functional constipation and its psychiatric comorbidities, researchers recruited a cohort of 187 patients. Participants were stratified using standardized self-reported assessments of depression and anxiety into two primary groups: those with functional constipation and anxiety or depression (FCAD) and those with functional constipation but no anxiety or depression (FCNAD). To enhance the precision of the subsequent machine learning analysis, the researchers categorized these individuals using high-confidence and low-confidence labels. This strategy allowed the model to first establish a baseline from patients with the most unambiguous clinical profiles before incorporating more complex, borderline cases. This approach is critical for clinicians to understand, as it mirrors the diagnostic challenge of distinguishing between primary gastrointestinal distress and symptoms heavily modulated by central nervous system states.

Machine Learning and Diagnostic Accuracy

The researchers extracted specific brain structural imaging features to train and refine a classification model using a stagewise training approach (a step-by-step method to refine a machine learning model by first training on high-confidence data before incorporating more nuanced cases). By focusing on the structural integrity of white matter and the volume of gray matter regions, the model was designed to identify the subtle neurological signatures that distinguish FCAD from FCNAD. The performance of this diagnostic tool was evaluated through rigorous cross-validation, where the classification model achieved an average accuracy of 89.23% in distinguishing between the two patient subtypes. Beyond simple binary classification, the model provided a predicted probability of a patient belonging to the FCAD group. This probability score demonstrated significant clinical utility, as it was significantly correlated with both mental ratings and the severity of gastrointestinal symptoms. For the practicing physician, these findings suggest that the degree of neurological alteration captured by the model reflects the objective burden of the disease, potentially offering a more precise method for identifying patients whose bowel symptoms are inextricably linked to their psychiatric status.

Neuroanatomical Markers of Comorbid Distress

The high diagnostic accuracy of the model was driven by a specific subset of neurological indicators, as the researchers identified the top 30 imaging features that most significantly contributed to the classification of patients with comorbid distress. These features were primarily concentrated in regions responsible for emotional processing, somatosensory perception, and motor control. Specifically, the model identified structural alterations in the temporal pole, the amygdala (a key center for processing fear and emotional memory), and the orbitofrontal cortex, which is a region involved in decision-making and the regulation of emotional responses. For the clinician, the involvement of these areas suggests that the psychiatric symptoms observed in these patients are rooted in measurable neuroanatomical changes rather than being merely secondary reactions to chronic physical discomfort. Beyond emotional centers, the study highlighted the role of the insula, a key somatosensory region that integrates internal bodily sensations with emotional states. This finding provides a neurological basis for the heightened visceral sensitivity often reported by patients with comorbid gastrointestinal and psychiatric distress. Furthermore, the classification model relied on structural data from regions governing motor control, including the corticospinal tract and the inferior cerebellar peduncle. The inclusion of these white matter tracts (the bundles of nerve fibers that facilitate communication between distant brain regions) indicates that the pathophysiology of functional constipation with comorbid distress involves a complex interplay between the brain's emotional circuitry and the descending pathways that regulate motor output.

Clinical Implications for Refractory Symptoms

The identification of distinct neuroanatomical markers addresses a critical challenge in gastroenterology, as gastrointestinal symptoms in patients with both functional constipation and anxiety or depression are often not fully resolved with standard medication. These patients frequently present with refractory symptoms that do not respond to conventional laxative therapies in the same manner as those without psychiatric comorbidities. The researchers noted that treatment responses in patients with comorbid distress differ significantly from those of patients without anxiety and depression, suggesting that a one size fits all approach to bowel management may be insufficient for this specific population. By demonstrating that variations in brain function and structure can serve as imaging features to differentiate between the two subtypes, the researchers have provided a biological basis for why some patients remain symptomatic despite adherence to traditional protocols. These findings highlight that brain imaging features may provide incremental value over standard behavioral assessments, such as self-reported anxiety and depression scales, which can be subject to patient bias or limited insight. In the future, integrating these objective neurological markers into clinical practice could facilitate more personalized diagnosis and treatment strategies, allowing clinicians to identify which patients require integrated neuro-gastroenterological interventions to achieve symptom resolution.

References

1. Chen Z, Peng Y, Shi Q, et al. Prevalence and Risk Factors of Functional Constipation According to the Rome Criteria in China: A Systematic Review and Meta-Analysis.. Frontiers in medicine. 2022. doi:10.3389/fmed.2022.815156

2. Huai Y, Fan Q, Dong Y, et al. Efficacy and mechanism of acupuncture for functional constipation in older adults: study protocol for a randomized controlled trial. Frontiers in Neurology. 2024. doi:10.3389/fneur.2024.1341861

3. Wang R, Xia M, Gao Z, Yu X, Guo A. Efficacy of Biofeedback Therapy on Chronic Constipation and Sexual Dysfunction in Postmenopausal Women: A Randomized Controlled Trial.. Neurogastroenterology and motility. 2025. doi:10.1111/nmo.70084

4. Xu S, Li J, Wang A. Electroacupuncture versus 5-HT4 receptor agonist for functional constipation: A systematic review and meta-analysis of randomized controlled trials. Medicine. 2024. doi:10.1097/MD.0000000000040634

5. Cryan JF, O’Riordan KJ, Cowan CS, et al. The Microbiota-Gut-Brain Axis. Physiological Reviews. 2019. doi:10.1152/physrev.00018.2018

6. Wang L, Luo X, Qing X, et al. Symptom effects and central mechanism of acupuncture in patients with functional gastrointestinal disorders: a systematic review based on fMRI studies.. BMC gastroenterology. 2024. doi:10.1186/s12876-024-03124-y

7. Zhang W, Hu Y, Wei J, et al. Multimodal brain imaging-based classification of functional constipation subtypes using machine learning.. Journal of affective disorders. 2026. doi:10.1016/j.jad.2026.121831