P0797 - Unsupervised Machine Learning Based Automatic Detection and Characterization of Prominent Bowel Sounds: Feasibility Study in Healthy Subjects

Keerthy Gopalakrishnan, MD¹, Gayathri Yerrapragada, MS², Avneet Kaur, MD³, Pratyusha Muddaloor, MD⁴, Divyanshi Sood, MD⁵, Anmol Mohan, MD⁶, Swetha Rapolu, MD⁷, Gianeshwaree Alias Rachna Panjwani, MD⁷, Rabiah Aslam Ansari, MD⁷, Nagmeh Asadimanesh, MD⁷, Shiva Sankari Karuppiah, MS⁷, Shivaram Poigai Arunachalam, PhD⁷
1The Wright Center for Graduate Medical Education, Scranton, PA; 2Mayo Clinic Health System, Jacksonville, FL; 3MedStar Union Memorial Hospital, Baltimore, MD; 4Lower Bucks Hospital, Bristol, PA; 5Parkview Medical Center, Pueblo, CO; 6Carle Foundation hospital, Urbana, IL; 7Mayo Foundation for Medical Education and Research, Jacksonville, FL

Introduction: Non-Invasive gut health monitoring poses a pressing clinical need, based on which we reported our previous work demonstrating a You Only Listen Once (YOLO) deep learning model for automatic detection of prominent bowel sounds (BS) i.e., phonoenterogram (PEG). The YOLO model was developed based on expert manual annotations of prominent BS that is cumbersome and subjective. The purpose of this work was to develop an unsupervised machine learning platform to automatically annotate and classify BS recorded from healthy participants towards adaptive YOLO system for real-time PEG monitoring.

Methods: 110 2-Minute PEGs sampled at 44.1 kHz were recorded using the EKO Duo® stethoscope from 8 healthy volunteers at two locations LUQ and RLQ after IRB approval. PEG signals were processed to detect prominent BS events using an unsupervised model-free approach based on short-time energy (STE) and zero-crossing rate (ZCR). Next, we extracted multi-dimensional acoustic features from each prominent event, including Mel-frequency cepstral coefficients (MFCCs), spectral centroid, bandwidth, zero-crossing rate, and duration. We then applied KMeans clustering algorithm to organize events into clinically relevant categories. Clustering outputs were evaluated using internal metrics such as Silhouette Score with KMeans (k=5) producing optimal separation (Silhouette ≈ 0.6).

Results: From the full dataset, 42,975 total audio segments were extracted. Among these, 6,314 were labeled as prominent bowel sound events. Clustering revealed five reproducible morphologies across subjects from visual inspection as below:

1. Single Burst – brief, isolated, high-energy impulse
2. Harmonic – sustained waveforms with regular oscillations
3. Multiple Burst – sequences of impulse bursts
4. Random Continuous – irregular high-frequency sequences
5. Multi-Modal – overlapping or complex compound patterns



KMeans with k=5 achieved optimal separation Cluster validation metrics supported the reproducibility of the five-class scheme shown in Figure 1A and 1B shows representative time-domain waveforms of bowel sound categories.

Discussion: This study presents an unsupervised and scalable framework for characterizing bowel sounds using real-world recordings demonstrating the feasibility for non-invasive, label-free acoustic profiling of gastrointestinal activity and can serve as a foundation for clinical GI monitoring applications for variety of bowel diseases.


Figure: Figure 1A. Representative Prominent Bowel Sounds from Each Cluster.
Figure 1B. Characteristic Waveforms of Five Bowel Sound Categories


Disclosures:
Keerthy Gopalakrishnan indicated no relevant financial relationships.
Gayathri Yerrapragada indicated no relevant financial relationships.
Avneet Kaur indicated no relevant financial relationships.
Pratyusha Muddaloor indicated no relevant financial relationships.
Divyanshi Sood indicated no relevant financial relationships.
Anmol Mohan indicated no relevant financial relationships.
Swetha Rapolu indicated no relevant financial relationships.
Gianeshwaree Alias Rachna Panjwani indicated no relevant financial relationships.
Rabiah Aslam Ansari indicated no relevant financial relationships.
Nagmeh Asadimanesh indicated no relevant financial relationships.
Shiva Sankari Karuppiah indicated no relevant financial relationships.
Shivaram Poigai Arunachalam indicated no relevant financial relationships.

Keerthy Gopalakrishnan, MD¹, Gayathri Yerrapragada, MS², Avneet Kaur, MD³, Pratyusha Muddaloor, MD⁴, Divyanshi Sood, MD⁵, Anmol Mohan, MD⁶, Swetha Rapolu, MD⁷, Gianeshwaree Alias Rachna Panjwani, MD⁷, Rabiah Aslam Ansari, MD⁷, Nagmeh Asadimanesh, MD⁷, Shiva Sankari Karuppiah, MS⁷, Shivaram Poigai Arunachalam, PhD⁷. P0797 - Unsupervised Machine Learning Based Automatic Detection and Characterization of Prominent Bowel Sounds: Feasibility Study in Healthy Subjects, ACG 2025 Annual Scientific Meeting Abstracts. Phoenix, AZ: American College of Gastroenterology.