Beyond Anatomy: How fMRI and DTI Help Map the Brain

A brain scan can reveal a tumor. It can show its size, location, and how it affects the surrounding tissue. But before surgery, another question becomes just as important: what does the brain tissue around that tumor actually do?

Imagine a tumor growing near the part of the brain responsible for movement. A conventional MRI may show the tumor clearly, but if it has distorted the surrounding anatomy, identifying the nearby motor-related regions becomes harder.

This is where functional neuroimaging becomes valuable.

In their 2015 review, Functional Neuroimaging: Fundamental Principles and Clinical Applications, Khanna and colleagues explore two complementary techniques: functional magnetic resonance imaging (fMRI) and diffusion tensor imaging (DTI).

One helps us investigate brain activity. The other provides information about the pathways connecting different brain regions.

Together, they allow us to look beyond what the brain looks like and begin examining how it works.

---

fMRI: Understanding brain activity through blood

When you move your fingers, speak, or process what you see, particular groups of neurons become active. These neurons need oxygen to support their activity, so the brain responds by increasing blood flow to the regions involved.

Interestingly, the increase in blood flow is usually greater than the increase in oxygen consumption. As a result, the local concentration of deoxygenated hemoglobin, or deoxyhemoglobin, decreases.

Why does this matter for MRI?

Deoxyhemoglobin slightly disturbs the magnetic field around it, causing the MRI signal to fade more quickly. This fading is described by a property called T2 star, or T2*.

When deoxyhemoglobin decreases, the magnetic disturbance decreases, so the signal fades more slowly and appears slightly stronger on a T2*-sensitive image.

This is the basis of blood-oxygen-level-dependent (BOLD) fMRI.

The relationship between neural activity and the accompanying changes in blood flow and oxygenation is called neurovascular coupling.

Importantly, fMRI does not directly record neurons firing. Instead, it measures a blood oxygenation-related response from which we infer aspects of neural activity.

How does this become a brain map?

Consider a patient asked to tap their fingers repeatedly while lying inside an MRI scanner.

The patient alternates between tapping and resting while the scanner continuously collects brain images.

Each image contains thousands of small three-dimensional units called voxels, essentially the 3D equivalent of pixels.

Researchers examine how the MRI signal changes over time in each voxel and determine whether it follows the expected response to the finger-tapping task.

The results are displayed as a colored activation map, highlighting regions whose signals are statistically associated with the task.

This can help surgeons identify motor-related cortex when a tumor has distorted the brain’s normal anatomical landmarks.

But the colors are statistical results, not photographs of neurons firing.

A missing response does not necessarily mean the underlying tissue has no function. For example, abnormal blood vessels near a tumor may fail to respond normally to neural activity, a phenomenon called neurovascular uncoupling.

The neurons may still be functioning, even though the expected BOLD response is weak or absent.

---

What happens when the brain is resting?

The brain does not become inactive when we stop performing a task.

Even at rest, different regions continue to show spontaneous fluctuations in their BOLD signals.

Interestingly, some regions show similar patterns. When the signal rises in one region, it may also rise in another, and when it falls, the other region may show a similar change.

Researchers describe this statistical relationship as functional connectivity.

Resting-state fMRI uses these patterns to identify organized networks of brain regions, including the default mode network, which is associated with internally oriented processes such as aspects of memory and self-referential thought.

However, functional connectivity does not necessarily mean that two regions are directly connected by nerve fibers.

To investigate the brain’s physical pathways, we need a different approach.

---

DTI: Following the movement of water

While fMRI investigates changes related to blood oxygenation, diffusion tensor imaging takes advantage of something happening continuously throughout the brain: the movement of water molecules.

Water molecules are constantly moving randomly, a process called Brownian motion.

In an environment where water can move equally in every direction, diffusion is described as isotropic.

But brain tissue contains structures that influence how water moves.

White matter, for example, is rich in axons, the long extensions of neurons that carry signals between different brain regions.

Within organized bundles of axons, water generally diffuses more easily along the length of the bundle than across it, partly because of barriers created by cellular structures.

This direction-dependent movement is called anisotropic diffusion.

DTI measures water diffusion in multiple directions, allowing us to estimate the predominant orientation of diffusion within brain tissue.

From water movement to white matter pathways

DTI uses a mathematical model called a tensor to summarize the diffusion pattern within each voxel.

One common measurement derived from this model is fractional anisotropy, or FA, which describes how strongly diffusion favors particular directions.

Computer algorithms can then use the estimated diffusion directions across neighboring voxels to reconstruct possible white matter pathways.

This process is called tractography.

In one case discussed by Khanna and colleagues, tractography shows how the corticospinal tract, a major pathway involved in movement, runs around a brain tumor.

Knowing the estimated position of this pathway can provide valuable information during surgical planning.

However, a single voxel may contain thousands of axons traveling in different directions. The computer cannot resolve every individual fiber, so it may miss a real pathway or reconstruct one that does not actually exist.

Tractography produces estimates of possible pathways, not photographs of individual nerve fibers.

---

Why both techniques matter

Returning to our original tumor example, we can now see why these techniques are useful together.

Structural MRI shows the tumor and surrounding anatomy. fMRI helps identify nearby regions associated with important functions, while DTI helps estimate the course of white matter pathways.

Combined with the patient’s neurological examination and other clinical evidence, these techniques can provide additional information for surgical planning.

The review also discusses applications in epilepsy, dementia, and multiple sclerosis, where changes in functional connectivity or water diffusion may provide information beyond conventional imaging. However, neither technique should be considered a standalone diagnostic test for these conditions.

Ultimately, the value of functional neuroimaging is not simply in producing colorful brain maps, but in revealing information that anatomy alone cannot provide.

And to interpret those maps correctly, we must always understand what was measured, what was inferred, and what the technique might have missed.

---

Source: Khanna et al., "Functional Neuroimaging: Fundamental Principles and Clinical Applications," The Neuroradiology Journal (2015).