AI for Neurology: Brain Imaging and Neurodegenerative Disease

Module: Healthcare AI | Difficulty: Advanced

Brain Age Gap

Voxel-Based Morphometry

Neurological AI Applications

import torch
import torch.nn as nn
import torchvision.models as models

class BrainAgePredictor(nn.Module):
    def __init__(self):
        super().__init__()
        self.encoder = models.resnet50(pretrained=True)
        num_features = self.encoder.fc.in_features
        self.encoder.fc = nn.Sequential(
            nn.Linear(num_features, 256), nn.ReLU(),
            nn.Dropout(0.3), nn.Linear(256, 1))

    def forward(self, x):
        return self.encoder(x)

class BrainSegmentation(nn.Module):
    def __init__(self, num_classes=4):
        super().__init__()
        self.encoder1 = self._conv_block(1, 32)
        self.encoder2 = self._conv_block(32, 64)
        self.encoder3 = self._conv_block(64, 128)
        self.pool = nn.MaxPool3d(2)
        self.bottleneck = self._conv_block(128, 256)
        self.up3 = nn.ConvTranspose3d(256, 128, 2, stride=2)
        self.dec3 = self._conv_block(256, 128)
        self.up2 = nn.ConvTranspose3d(128, 64, 2, stride=2)
        self.dec2 = self._conv_block(128, 64)
        self.up1 = nn.ConvTranspose3d(64, 32, 2, stride=2)
        self.dec1 = self._conv_block(64, 32)
        self.out = nn.Conv3d(32, num_classes, 1)

    def _conv_block(self, in_ch, out_ch):
        return nn.Sequential(
            nn.Conv3d(in_ch, out_ch, 3, padding=1),
            nn.InstanceNorm3d(out_ch), nn.LeakyReLU(0.1),
            nn.Conv3d(out_ch, out_ch, 3, padding=1),
            nn.InstanceNorm3d(out_ch), nn.LeakyReLU(0.1))

    def forward(self, x):
        e1 = self.encoder1(x)
        e2 = self.encoder2(self.pool(e1))
        e3 = self.encoder3(self.pool(e2))
        b = self.bottleneck(self.pool(e3))
        d3 = self.dec3(torch.cat([self.up3(b), e3], 1))
        d2 = self.dec2(torch.cat([self.up2(d3), e2], 1))
        d1 = self.dec1(torch.cat([self.up1(d2), e1], 1))
        return self.out(d1)

model = BrainAgePredictor()
mri = torch.randn(1, 3, 224, 224)
predicted_age = model(mri)
print(f'Predicted brain age: {predicted_age.item():.1f}')

seg_model = BrainSegmentation(num_classes=4)
vol = torch.randn(1, 1, 128, 128, 128)
seg = seg_model(vol)
print(f'Brain segmentation output: {seg.shape}')

Research Insight: Brain age prediction from MRI provides a biomarker for neurodegenerative disease: individuals whose brain age exceeds their chronological age are at higher risk for Alzheimer's and other dementias. The brain age gap can detect preclinical disease 5-10 years before clinical symptoms, but requires careful harmonization across sites.