Harnessing spatial MRI normalization: patch individual filter layers for CNNs

TL;DR

This paper introduces patch individual filter (PIF) layers for CNNs, leveraging spatial normalization in MRI to improve brain image analysis, especially with limited data, by training local filters that capture region-specific features.

Contribution

The paper proposes a novel PIF layer type that trains local filters in CNNs, exploiting MRI spatial normalization to enhance performance in neuroimaging tasks.

Findings

PIF layers outperform standard CNNs in multiple neuroimaging tasks.

PIF layers show significant benefits in low sample size scenarios.

The approach improves accuracy in sex, AD, and MS classification.

Abstract

Neuroimaging studies based on magnetic resonance imaging (MRI) typically employ rigorous forms of preprocessing. Images are spatially normalized to a standard template using linear and non-linear transformations. Thus, one can assume that a patch at location (x, y, height, width) contains the same brain region across the entire data set. Most analyses applied on brain MRI using convolutional neural networks (CNNs) ignore this distinction from natural images. Here, we suggest a new layer type called patch individual filter (PIF) layer, which trains higher-level filters locally as we assume that more abstract features are locally specific after spatial normalization. We evaluate PIF layers on three different tasks, namely sex classification as well as either Alzheimer's disease (AD) or multiple sclerosis (MS) detection. We demonstrate that CNNs using PIF layers outperform their counterparts in several, especially low sample size settings.