US-12622623-B1 - System and method for predictive modeling and analysis of neuron flow

US12622623B1US 12622623 B1US12622623 B1US 12622623B1US-12622623-B1

Abstract

A neuron pulse acquisition system is disclosed for capturing and processing neuron flow signals from the human body in a non-invasive manner. The system includes a pad with a skin sensor configured to detect neuron pulse potentials, a variable gain amplifier (VGA) to amplify detected signals, and an analog filter (A-FILTER) to remove noise. An analog-to-digital converter (ADC) digitizes the filtered signals, which are processed by a digital signal processing (DSP) and Bluetooth unit executing Fourier Transform and Cross-Correlation algorithms to analyze neuron flow characteristics in time and frequency domains. A D/A calibration module maintains analog accuracy, and a system software controller performs spectral analysis, cross-correlation, and machine learning-based predictive modeling. The system enables real-time monitoring, classification, and diagnostic interpretation of neuron flow behavior, thereby facilitating predictive analysis of neural conditions with high precision and reliability.

Inventors

Saf Asghar
James Vero Asghar
Miki Moyal

Assignees

Saf Asghar
James Vero Asghar
Miki Moyal

Dates

Publication Date: 20260512
Application Date: 20251118

Claims (8)

1 . A neuron pulse acquisition system for capturing and processing neuron flow signals from a human body, comprising: a pad configured to be adhesively attached on skin of a subject for non-invasive acquisition of neuron pulse signals; a skin sensor disposed within the pad, the skin sensor configured to detect electrical potentials generated by neuron pulses from underlying skin surface of the subject; a Variable Gain Amplifier (VGA) coupled to the skin sensor, the VGA configured to amplify the detected neuron pulse signals within a predetermined millivolt range corresponding to neuron pulse activity; an Analog Filter (A-FILTER) coupled to the VGA, the analog filter configured to remove unwanted noise components from the amplified neuron pulse signals; an Analog-to-Digital Converter (ADC) configured to digitize the filtered neuron pulse signals, the ADC having a resolution of at least 18-bits and a dynamic range below one millivolt; a Digital Signal Processing (DSP) configured to: process the digitized neuron pulse signals using at least one of a Fourier Transform or a Cross-Correlation algorithm to obtain neuron flow characteristics in at least one of a time domain or a frequency domain; a bluetooth unit communicatively coupled to the ADC, configured to: wirelessly transmit processed neuron flow data to an external computing device; a D/A calibration module configured to provide feedback calibration for analog sections of the system to maintain signal accuracy; and a system software controller executed by the DSP configured to: autonomously execute signal processing routines for spectral analysis and predictive modeling of neuron flow characteristics; perform cross-correlation analysis between neuron pulse profiles to identify asymmetries indicative of abnormal neuron flow characteristics; and classifies neuron pulse profiles into normal and abnormal categories based on cross-correlation coefficients and generates corresponding diagnostic indicators.
2 . The neuron pulse acquisition system of claim 1 , wherein the pad comprises a self-adhesive skin-contact surface configured to provide stable electrode coupling and includes noise cancelation circuitry integrated within the skin sensor to suppress ambient and physiological interference.
3 . The neuron pulse acquisition system of claim 1 , wherein the ADC comprises an oversampling sigma-delta converter designed for mixed-signal integration on a single silicon substrate with the analog front-end circuitry to enable low-power operation.
4 . The neuron pulse acquisition system of claim 1 , wherein the VGA is configured to dynamically adjust gain based on detected neuron pulse intensity to maintain an output voltage within a range of 0 to 10 millivolts corresponding to physiological neuron potential limits.
5 . The neuron pulse acquisition system of claim 1 , wherein the DSP and Bluetooth unit comprises a multiply-accumulate (MAC) subsystem and an embedded non-volatile memory for storing neuron pulse datasets and executing Fast Fourier Transform (FFT) algorithms with at least 1024 bin sizes for spectral analysis.
6 . The neuron pulse acquisition system of claim 1 , wherein the system software controller includes machine learning algorithms configured to autonomously generate and update a library of neuron pulse profiles, each representing a different physiological or pathological condition, wherein the machine learning algorithms utilize neural network or statistical learning models to adaptively refine predictive parameters for identifying deviations in neuron flow patterns.
7 . The neuron pulse acquisition system of claim 1 , wherein the pad is further configured to harvest electrical current from the body of the subject for capturing and processing neuron flow signals.
8 . The neuron pulse acquisition system of claim 1 , wherein the D/A calibration module is configured to apply feedback control to the analog front-end components, including the VGA and A-FILTER, thereby compensating for drift, temperature variation, and signal distortion during long-duration operation.

Description

FIELD OF THE INVENTION The present disclosure generally relates to the field of biomedical signal acquisition and analysis, and more particularly, to a neuron pulse acquisition system and method configured for capturing and processing neuron flow signals from the human body. BACKGROUND OF THE INVENTION The human nervous system comprises a complex network of neurons responsible for transmitting electrical impulses between the brain and various parts of the body. These electrical impulses, often referred to as neuron pulses or action potentials, are critical in enabling sensory perception, motor control, and cognitive function. Monitoring and analyzing neuron flow provides valuable insight into neural communication patterns and potential abnormalities associated with neurological disorders. Conventional systems for measuring neuronal activity rely predominantly on invasive electrodes or surface-based electroencephalography (EEG) and electromyography (EMG) techniques. While such approaches allow detection of neuron activity, they are often limited in their ability to resolve fine-grained signal characteristics, particularly those associated with localized neuron pulse propagation. Moreover, these methods are generally constrained to time-domain analyses, providing limited visibility into frequency-domain spectral content that could reveal deeper insights into neuron flow behavior. Existing analog front-end circuits used in biomedical sensors often suffer from low dynamic range, amplifier drift, and high susceptibility to electrical noise, thereby reducing the accuracy of captured neuron signals. Furthermore, the lack of integration between analog and digital processing architectures increases design complexity and power consumption, which restricts their use in portable or wearable medical devices. Although advancements in digital signal processing (DSP) and machine learning have improved data interpretation in certain biomedical contexts, their application to neuron pulse modeling remains limited. Traditional systems rarely employ spectral-domain analysis, such as Fast Fourier Transform (FFT) or Cross-Correlation techniques, for identifying neuron flow abnormalities. Likewise, there is an absence of self-learning predictive models that can autonomously classify neuron flow patterns into normal and abnormal categories. Additionally, the conventional systems often rely on external power sources or rechargeable batteries, which increase overall bulk and limit continuous, long-term operation required for practical wearable deployment. Accordingly, there exists a need for a cost-effective, energy-efficient, non-invasive, and integrated system capable of capturing, digitizing, and analyzing neuron pulse signals with enhanced accuracy and stability. There is also a need for such a system to employ frequency-domain spectral analysis and artificial intelligence-based predictive modeling to identify deviations in neuron flow profiles for early diagnosis and neurological assessment. SUMMARY OF THE INVENTION The present disclosure provides a neuron pulse acquisition system and method for capturing, digitizing, and analyzing neuron flow signals in the human nervous system. The invention is directed toward the creation of a statistical and predictive model of neuron flow or propagation within neural transport structures of the body. Understanding neuron flow is a necessary conduit to comprehending the electrical behavior of the nervous system under various physiological and pathological conditions. The disclosure further considers the integration of artificial intelligence (AI) and machine learning (ML) techniques for predictive modeling of neuron flow behavior, including any departures from normal patterns. From empirical and publicly available data, the inventors have identified measurable indicators of neuron flow characteristics using existing signal processing algorithms, thereby motivating the development of a dedicated, non-invasive acquisition and processing system. With the evolution of digital signal processing (DSP) capabilities and computing performance, it has become possible to conduct detailed analysis of neuron flow behavior in the frequency domain, offering significant advantages over conventional time-domain-only analysis. Traditional time-domain observations of action potentials do not adequately characterize the internal content or spectral composition of neuron transport profiles. Frequency-domain techniques, such as Discrete Fourier Transform (DFT) and Fast Fourier Transform (FFT), enable the extraction of richer information regarding the spectral behavior of neuron signals. The disclosed system therefore employs an integrated analog and digital signal processing architecture, including a non-invasive adhesive pad, Variable Gain Amplifier (VGA), analog filter, Analog-to-Digital Converter (ADC), and Digital Signal Processing (DSP) module, all operating cohesively to acquire neuron pulses with high prec