Optical Measurement Noise Removal Using Random Matrix Theory
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical measurement systems face challenges in accurately detecting neural activity in the brain due to noise introduced by layers of tissue and components of the measurement system, such as the scalp and skull, as well as defects and impurities in optical components.
Innovation Solution
The implementation of a processing unit that applies random matrix theory to model noise in histogram data, allowing for the filtration of noise and generation of clean histogram data, which represents the meaningful neural activity response.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If light pulses are used to detect neural activity through tissue layers, then neural activity can be measured, but noise is introduced by tissue layers and optical components
Solution Approach 1:
The patent segments the histogram data into signal components and noise components using random matrix theory. The total histogram is divided into a signal subspace (containing neural activity information) and a noise subspace (containing artifacts from tissue layers and optical components), allowing selective processing of each component.
Solution Approach 2:
The patent extracts the noise component from the total histogram data by identifying and removing contributions from tissue layers (scalp, skull, CSF) and optical component defects. This extraction process isolates the pure neural activity signal from confounding factors.
2Loss of information
If histogram data is collected from all photon arrivals, then complete neural activity information is captured, but data dimensionality and processing requirements increase
Solution Approach 1:
The patent extracts only the relevant signal components from the complete histogram data using random matrix theory. By identifying the signal subspace, the system extracts neural activity information while discarding redundant noise components, reducing data dimensionality while preserving essential information.
Solution Approach 2:
The patent transforms the high-dimensional histogram data into a lower-dimensional representation by changing the parameter space from all photon arrival times to only those within the signal subspace. This parameter transformation reduces computational complexity while maintaining measurement accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach results in more accurate and useful metrics for neural activity analysis, reducing resource requirements for processing, transmission, and storage by providing clean histogram data of reduced dimensionality.
Implementation Method 1
a light source configured to emit light directed at a target within a user
Implementation Method 2
detect photon arrival times for photons of the light after the light is scattered by the target
Data Source
AI summary
An illustrative optical measurement system includes a light source configured to emit light directed at a target within a user. The system further includes a detector configured to detect photon arrival times for photons of the light after the light is scattered by the target. The system further includes a processor configured to determine, based on the photon arrival times, histogram data associated with the target, the histogram data including noise. The processor is further configured to determine, based on the photon arrival times, a random matrix corresponding to the photon arrival times. The processor is further configured to determine, based on the random matrix, a noise distribution representing a distribution of the noise within the histogram data. The processor is further configured to generate clean histogram data using the noise distribution to filter at least a portion of the noise from the histogram data.