Neural network based forward prediction of bladder pressure using pudendal nerve electrical activity

Area of Science:

• Biomedical Engineering
• Neuroscience
• Urology

Background:

• Spinal cord injury and neurological disorders impair bladder storage and voiding, leading to detrusor overactivity and incontinence.
• Current bladder control strategies aim to increase bladder capacity via electrical nerve stimulation, requiring continuous bladder pressure monitoring.

Purpose of the Study:

• To propose and evaluate a real-time bladder pressure estimation method using artificial neural networks.
• To assess the accuracy and predictive capabilities of the proposed method for bladder pressure.

Main Methods:

• Utilized artificial neural networks (ANNs) to estimate bladder pressure.
• Employed electroneurogram (ENG) signals from the pudendal nerve as input data.
• Conducted experiments on three rats to validate the real-time estimation approach.

Main Results:

• The ANN model accurately estimated synthetic bladder pressure during bladder contractions.
• The neural predictor demonstrated good generalization ability for bladder pressure prediction.
• Achieved average errors of 9.62% for 1-second ahead and 10.54% for 5-second ahead predictions.

Conclusions:

• Artificial neural networks offer a viable method for real-time bladder pressure estimation.
• This approach shows promise for enhancing bladder management and improving continence in patients with neurological dysfunction.
• Accurate bladder pressure prediction can support advanced neuroprosthetic control for bladder function.