NeuroQuant is FDA-cleared software that automatically measures the size of brain structures from an MRI scan. It takes a standard brain MRI, segments it into dozens of anatomical regions, and compares each region’s volume to a normative database of healthy brains matched by age and sex. The result is a report that tells a doctor whether specific parts of your brain are larger or smaller than expected, which can help in evaluating conditions like Alzheimer’s disease, epilepsy, traumatic brain injury, and multiple sclerosis.
How NeuroQuant Works
NeuroQuant is a post-processing tool, meaning it doesn’t require a special type of MRI. It works with a standard 3D brain scan taken on most modern MRI machines, including scanners with field strengths as low as 1.2 Tesla. Once the scan is complete, the raw images are sent to the NeuroQuant system for analysis.
The software first corrects for technical artifacts that vary from one scanner to another, such as distortions from magnetic field gradients and uneven signal intensity across the image. This step is important because it means results can be meaningfully compared across different MRI machines and over time. After correction, the scan is aligned to an internal anatomical atlas using a series of spatial transformations. The software then uses probabilistic methods and neural network models to label each tiny cube of brain tissue (called a voxel) with the anatomical structure it belongs to, based on its location and signal characteristics.
The entire process is automated. A radiologist or neurologist doesn’t need to manually trace brain regions, which removes the subjectivity and time cost of hand-drawn measurements. The output includes color-coded segmented images and volumetric reports showing how each brain region compares to the normative reference population.
What It Measures
NeuroQuant evaluates over 20 brain regions in a single analysis. The structures most relevant to clinical decision-making include the hippocampus (critical for memory and one of the first areas to shrink in Alzheimer’s disease), the lateral ventricles (fluid-filled spaces that enlarge as surrounding brain tissue atrophies), and the temporal horn of the ventricles, which sits right next to the hippocampus and can signal early volume loss there.
Beyond these key regions, the software also measures total brain volume, cortical gray matter, white matter, the amygdala, the thalamus, the caudate, the putamen, the cerebellum, and other structures. When a T2 FLAIR MRI sequence is also provided, NeuroQuant can quantify white matter lesions, the bright spots that commonly appear in multiple sclerosis or small vessel disease. A separate sequence can detect microbleeds, which are relevant in traumatic brain injury and certain types of dementia.
Validation studies comparing NeuroQuant to FreeSurfer, a widely used research-grade segmentation tool, have shown good to excellent agreement across most of the regions both tools measure. NeuroQuant’s overall volume measurements typically fall within about 2% of expert reference standards, with similar accuracy whether the patient has Alzheimer’s disease, mild cognitive impairment, or no neurological condition at all.
Clinical Uses
The most common reason a doctor orders NeuroQuant is to help evaluate memory loss. In early Alzheimer’s disease, the hippocampus begins shrinking years before symptoms become obvious. A standard MRI might look “normal” to the naked eye, but NeuroQuant can flag that the hippocampus is, say, at the 5th percentile for a person of that age, providing objective evidence that something is happening. Tracking these volumes over repeat scans also lets doctors monitor the rate of change, which can be more informative than any single measurement.
In epilepsy, NeuroQuant helps identify asymmetry between the left and right hippocampus. A smaller hippocampus on one side can point to mesial temporal sclerosis, one of the most common structural causes of seizures. Spotting this asymmetry strengthens the case for surgery when medications aren’t controlling seizures effectively.
For traumatic brain injury, including concussion, the software can detect subtle overall brain volume loss or enlargement of the ventricles that might not be apparent on a visual read of the scan. This is particularly useful when a patient reports persistent symptoms but conventional imaging looks unremarkable.
In multiple sclerosis, quantifying lesion burden and tracking brain atrophy over time gives neurologists a more precise way to gauge disease activity and treatment response than relying on visual impressions alone.
What the Report Looks Like
NeuroQuant generates several report types depending on the clinical question. The most widely used is the Age-Related Atrophy Report, which plots the volume of key structures against a normative curve. You’ll see a chart for each brain region with your measurement shown as a dot on a graph. If that dot falls within the middle range, the volume is typical for your age. If it falls near the bottom, it indicates the structure is smaller than expected.
Reports also include a color-coded overlay on the MRI itself, showing which voxels were assigned to which structure. This lets the ordering physician visually verify that the segmentation looks reasonable and wasn’t thrown off by motion artifacts or unusual anatomy.
The numbers are presented as both absolute volumes in cubic centimeters and as percentiles relative to the normative database. Percentiles are generally more useful because brain size varies considerably between individuals. What matters clinically is not the raw number but where you fall compared to others your age.
Limitations to Keep in Mind
NeuroQuant is a decision-support tool, not a diagnostic test. A low hippocampal volume doesn’t by itself mean you have Alzheimer’s disease. Brain volumes vary naturally, and shrinkage can result from many causes, including depression, alcohol use, sleep apnea, and normal aging. The results need to be interpreted alongside clinical symptoms, cognitive testing, and sometimes additional biomarkers.
Image quality matters significantly. Patient movement during the scan, improper scanner settings, or image compression can all degrade the segmentation accuracy. The manufacturer provides detailed scanner-specific protocols, and deviating from them can reduce the reliability of the output. If a scan is too noisy or has too much motion artifact, the software may produce inaccurate results or fail to process entirely.
The normative database also has boundaries. It is built from a specific population, and individuals whose brain anatomy falls outside typical ranges for reasons unrelated to disease (such as prior brain surgery or significant congenital variants) may get misleading percentile rankings. Radiologists reviewing the report are expected to account for these possibilities.
How It Differs From a Standard MRI Read
A conventional MRI report is based on a radiologist’s visual assessment. They look for obvious abnormalities: tumors, strokes, large areas of atrophy, lesions. This approach is excellent for detecting focal, dramatic changes but less reliable for subtle, diffuse volume loss. Two radiologists looking at the same scan might disagree on whether the hippocampus appears “mildly atrophied” or “within normal limits.”
NeuroQuant replaces that subjective call with a number. It won’t catch everything a trained radiologist’s eye will, such as a small tumor or a vascular malformation, but it excels at precisely quantifying volume in a way human perception cannot. The two approaches are complementary. Most clinicians use NeuroQuant alongside, not instead of, a standard radiological interpretation.