Development and Validation of Delirium Recognition Using Computer Vision in Neuro-critical Patients

Recruiting

• Conditions: Delirium; Artificial Intelligence (AI)

• Registration Number: NCT07136207
• Lead Sponsor: Beijing Tiantan Hospital

Brief Summary

This research project employs machine learning algorithms integrated with computer vision, image processing, and pattern recognition technologies to perform digital analysis of facial expression behaviors in neurocritical care patients with delirium. By constructing multidimensional high-level features of delirium, the investigators have established a classification model based on behavioral. The primary objective of this study is to address the critical challenge of achieving precise and efficient delirium diagnosis in neurologically critically ill patients through automated facial expression behavior recognition.

Detailed Description

This study is a prospective cohort study approved by the Ethics Committee of Beijing Tiantan Hospital. It aims to support the accurate and efficient diagnosis of delirium in neurocritical patients through a facial expression recognition system. A mobile application was developed for this study, collaboratively designed by senior clinicians and engineers from the Institute of Computing Technology, Chinese Academy of Sciences. The application is based on a stimulus paradigm designed using CAM-ICU (Confusion Assessment Method for the Intensive Care Unit) questions to record dynamic facial videos of neurocritical patients following delirium evaluation based on the DSM-V criteria.

Patients were assessed for delirium and facial expression behavior data were collected twice daily during ICU admission, in two time slots: 8:00-10:00 AM and 8:00-10:00 PM, following the study's inclusion and exclusion criteria. A trained and experienced specialist used the gold standard DSM-V to diagnose delirium. Within five minutes after completing the assessment, dynamic facial behavior video data were collected to prepare images for subsequent model development.

Various image preprocessing and data augmentation techniques were employed to prepare the images for the VGG16 model. These techniques are standard for running convolutional neural network (CNN) models. Using the "preprocess_input"function from the Keras VGGFace module, the investigators standardized image color and size to ensure that each image met the expected input requirements for model training. For data augmentation, the investigators applied TensorFlow's "ImageDataGenerator" function to perform horizontal flipping, rotation, scaling, width and height shifting, and shearing. These augmentation techniques created a more diverse dataset, helping to prevent overfitting and improving the model's generalizability to new faces.

The investigators developed a binary classification model to identify delirium using a CNN with a pretrained backbone. The VGG16 model, based on deep learning, was adopted, leveraging transfer learning from VGGFace2, which possesses pre-existing facial feature recognition capabilities. Transfer learning allowed us to utilize prior knowledge to detect features more quickly, accurately, and with lower computational cost. The VGGFace2 model was employed for training.

Model performance was evaluated through internal validation at Beijing Tiantan Hospital and external validation at Guiyang Second People's Hospital, with metrics including accuracy, sensitivity, specificity, and F1 score. Additionally, to address the "black box" issue of machine learning, occlusion heatmap techniques were used to identify the most critical facial regions for delirium assessment, with the results visualized on a virtual face.

This model aims to support precise and efficient identification of delirium in neurocritical care units.

Study Timeline

• Study First Submitted: 2025-07-18
• Status Verified: 2025-08
• Study First Submitted QC: 2025-08-14
• Last Update Submitted: 2025-08-14
• Study First Posted: 2025-08-22
• Last Update Posted: 2025-08-22
• Study Start: 2025-08-30
• Primary Completion: 2025-12-30
• Study Completion: 2026-01-30

Recruitment & Eligibility

• Status: RECRUITING
• Sex: All
• Target Recruitment: 1000

Inclusion Criteria

1. Neurocritical patients admitted to the ICU, including postoperative neurosurgical patients, stroke patients, and those receiving ICU care due to other neurological conditions.
2. Age over 18 years.
3. Signed informed consent.

Exclusion Criteria

1. Age under 18 years.
2. Persistent coma (GCS ≤ 8) within 7 days pre- and post-surgery, making delirium assessment impossible.
3. Did not survive more than 24 hours in the ICU.
4. Patients with facial paralysis, post-traumatic facial disfigurement, or other conditions that could significantly affect facial recognition.
5. Exclusion of patients with severe dementia, Parkinson's disease, depression, or other conditions that might impact facial emotional expressions.

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
Accuracy of the delirium prediction model	Through study completion, an average of 1 year	The accuracy of the delirium prediction model will be calculated as the proportion of correct predictions among total predictions.
Sensitivity of the delirium prediction model	Through study completion, an average of 1 year	Sensitivity (true positive rate) will be assessed as the proportion of actual delirium cases correctly identified by the model.
Specificity of the delirium prediction model	Through study completion, an average of 1 year	Specificity (true negative rate) will be calculated as the proportion of non-delirium cases correctly identified by the model.

Secondary Outcome Measures

Name	Time	Method
F1 Score of the delirium prediction model	Through study completion, an average of 1 year	The F1 score, the harmonic mean of precision and recall, will be used to evaluate the balance between sensitivity and specificity.
AUC of the facial feature curve for delirium patients	Through study completion, an average of 1 year	The area under the curve (AUC) of the receiver operating characteristic (ROC) curve derived from facial features will be used to assess the discriminatory performance of the model.

Trial Locations

Locations (1)

Beijing Tiantan Hospital; 🇨🇳 Beijing, Beijing, China