# Machine learning-enhanced multi-band metamaterial sensor for early detection of neurological disorders

## TL;DR

This paper introduces a metamaterial sensor enhanced with machine learning for detecting early neurological disorders with high accuracy and efficiency.

## Contribution

A machine learning-enhanced metamaterial sensor with multi-band absorption for early neurological disorder detection is proposed.

## Key findings

- The sensor achieves over 99% absorption at three THz frequencies with high sensitivity.
- Machine learning reduces simulation time by 60% while maintaining accuracy.
- The sensor shows excellent performance across various brain tissues.

## Abstract

This study presents a square enclosed double octagonal shaped metamaterial absorber for the detection of initial neurological conditions. The size is very compact (\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$0.54 \lambda _0 \times 0.54 \lambda _0 \times 0.05 \lambda _0$$\end{document}) and achieves an absorption value over 99% for frequencies of 4.632 THz, 6 THz, and 7.384 THz. The key focus of the study is sensitivity, which is excellent for all of the peaks, which are 1.5 THz/RIU, 1.5 THz/RIU, and 1.8 THz/RIU, indicating its superior performance across multiple frequency bands. Furthermore, the q factor values are 22.21, 29.41, and 33.81, and the fom values are 7.19, 7.35, and 8.24 for the three frequency peaks, respectively, supporting its reliability in sensing. The electric field, magnetic field, and surface current are analyzed, and an equivalent circuit is created and evaluated for design validation. Additionally, machine learning techniques were employed to optimize and predict the structure’s performance, with the GradientBoosting model demonstrating a reduction of up to 60% in simulation time while maintaining high predictive accuracy. The overall performance and sensing tests across various brain tissues demonstrate excellent results, indicating that this sensor may be an ideal choice for the detection of early-stage neurological disorders.