Closed-Loop Deep Brain Stimulation Using ANN Sleep-State Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Traditional deep brain stimulation (DBS) systems are open-loop and non-adaptive, leading to inefficient energy consumption and battery depletion due to continuous stimulation regardless of the patient's activity, and they do not address sleep-related symptoms in Parkinson's disease.

Innovation Solution

Implementing a closed-loop DBS system that uses artificial neural networks to infer sleep stages from local field potentials, allowing for adaptive stimulation based on real-time brain activity, thereby optimizing therapy and conserving battery life.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional open-loop DBS systems provide continuous stimulation regardless of patient state, then motor symptoms are treated consistently, but energy consumption increases and battery life decreases

Engineering Contradiction:

Improveconsistent motor symptom treatmentVSAvoidenergy consumption

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The DBS system transitions from static continuous stimulation to dynamic adaptive stimulation that adjusts parameters based on real-time detection of patient physiological states (awake, asleep, moving, stationary) using machine learning algorithms analyzing local field potentials and other biosignals

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system implements closed-loop feedback by continuously monitoring patient brain activity through local field potentials, processing these signals through trained machine learning models to determine current state, and adjusting stimulation parameters accordingly to maintain therapeutic effectiveness while conserving energy

Inventive Principle:

Principle #23Feedback

2Duration of action of moving object

If traditional DBS systems operate continuously, then therapeutic coverage is maintained, but battery depletion occurs faster requiring more frequent replacements

Engineering Contradiction:

Improvetherapeutic coverage durationVSAvoidbattery life

Core Design Contradiction:

Duration of action of moving objectVSDuration of action of stationary object

Solution Approach 1:

The system employs periodic stimulation patterns adjusted to patient needs rather than continuous stimulation, delivering therapeutic effects during awake and active states while suspending or reducing stimulation during sleep periods, thereby extending battery operational duration

Inventive Principle:

Principle #19Periodic action

3Reliability

If DBS systems provide constant stimulation, then motor symptoms are managed, but sleep-related symptoms in Parkinson's disease are not addressed

Engineering Contradiction:

Improvemotor symptom managementVSAvoidsleep symptom treatment capability

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The DBS system is enhanced with multi-functionality to address both motor symptoms and sleep-related symptoms of Parkinson's disease by detecting sleep stages through machine learning analysis of local field potentials and adapting stimulation parameters to treat different symptom types appropriately

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Device complexity

If open-loop DBS systems are used, then system complexity is low, but adaptability to patient activity states is poor

Engineering Contradiction:

Improvesystem complexityVSAvoidadaptability to patient activity

Core Design Contradiction:

Device complexityVSAdaptability or versatility

Solution Approach 1:

Machine learning algorithms serve as intermediary components that process raw local field potential signals and translate them into state-specific stimulation commands, enabling adaptive behavior without requiring complex hardwired logic circuits in the implanted device

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS12539420B2Deep brain stimulation using artificial neural networks

Publication Date: 2026.02.03 THE REGENTS OF THE UNIVERSITY OF COLORADO

AI summary

Various embodiments of the present technology generally relate to closed loop deep brain stimulation based on inferred sleep stage from physiological data using machine learning classifiers. Some embodiments, for example, may use subthalamic nucleus (STN) deep brain stimulation (DBS) to treat advanced Parkinson's Disease motor symptoms and improve sleep by identifying sleep stages commensurate with clinician-scored polysomnography (PSG). The DBS may be adapted to include a novel artificial neural network (ANN) that triggers targeted stimulation in response to inferred sleep state from STN local field potentials (LFPs) recorded from implanted DBS electrodes. A feedforward neural network can be trained to prospectively identify sleep stage with PSG-level accuracy. In some embodiments, the machine learning model stored within the DBS may also adapt stimulation during specific sleep stages to treat targeted sleep deficits.