Structural Brain MRI of Mr. Abdou Traya: A Direct-from-DICOM Technical Review of a 2022 Clinical Study

1.Introduction

In May 2022, Mr. Abdou Traya underwent a clinically-ordered brain MRI at Calgary Foothills Medical Centre under the care of Alberta Health Services. The study was performed on a 1.5 Tesla GE Optima MR450w scanner as a standard structural examination to evaluate for organic neurological pathology. The complete DICOM archive of the study was retained by Mr. Traya, and he has subsequently reviewed it himself, including through consumer large-language-model interfaces. Those informal reviews have circulated various quantitative claims that are worth placing alongside what the dataset actually contains. This technical review does that, working directly from the DICOM binaries rather than from any downstream secondary document.

2.Methods

The DICOM archive (594 instances across 14 series) was read into Python 3.12 using pydicom v3.x. Header metadata was extracted from representative instances of each series. Pixel data were decompressed (JPEG Lossless, Process 14) and rendered to 8-bit grayscale using 1st–99th percentile window-levelling for qualitative review. Outer head dimensions were computed by thresholded bounding-box analysis on the axial T2 series, using the stated in-plane pixel spacing (0.4297 mm) and the actual inter-slice spacing derived from the ImagePositionPatient DICOM tag (2.92 mm). No automated brain parenchymal segmentation, registration, or regional volumetry was performed; the rationale for this restraint is given in Section 6.

3.Acquisition Protocol

This is a standard structural brain protocol designed for lesion detection: T1 for anatomy, T2 for parenchymal signal, FLAIR for white-matter and periventricular pathology, DWI with ADC for ischaemia, and SWI for microhaemorrhage and venous pathology. No 3D high-resolution isotropic T1 acquisition (MPRAGE, BRAVO, IR-SPGR, or equivalent) was acquired; this is a routine choice for clinical brain MRI and has no bearing on diagnostic adequacy, but it does constrain what can be derived downstream (Section 6).

4.Direct Measurements from the Archive

The following values were computed directly from the DICOM pixel data using the spacing metadata embedded in each file. They are presented with their method of derivation so that any reader can reproduce them.

These are outer head dimensions — they include scalp, subcutaneous tissue, and the skull vault. They do not represent intracranial volume (ICV) or brain parenchymal volume, which would require different sequences and segmentation methods (see Section 6). The SI extent is reported as a lower bound because it reflects the acquired volume; if the scan did not reach the absolute vertex of the skull, the true SI head dimension would be slightly greater than 148.9 mm.

For comparative reference, published anthropometric data for adult North American males give mean head length (glabella-opisthocranion, AP) of approximately 195–200 mm and mean head breadth (euryon-euryon, LR) of approximately 152–157 mm.^1,2 The measurements here — AP 215 mm, LR 184 mm — are above those means. The scan requisition itself contains the ordering physician's contemporaneous note describing a large head. The direct DICOM measurement reported here independently and objectively supports that observation.

5.Qualitative Review of Representative Images

Representative mid-slice images from four primary sequences are shown in Figure 1, rendered directly from the DICOM archive at 1st–99th percentile window-levelling with no further enhancement. These are shown for transparency and are not a substitute for the formal radiological report, which was rendered by the board-certified neuroradiologist of record at Alberta Health Services and which governs all clinical interpretation.

Across the representative slices reviewed, the study discloses no gross structural abnormality: ventricular system appears symmetric and of normal calibre, gray-white differentiation is preserved, no mass lesion or midline shift, no extra-axial collection, no acute infarct on DWI/ADC pairing, and no confluent white-matter disease. The expected clinical yield of a negative structural MRI in similar contexts is high — published series of first-episode MRI imaging in this setting report incidental or clinically-relevant findings in only a minority of cases.^4,5

6.Methodological Scope of This Dataset

This study was designed for clinical lesion detection, and it performs that task appropriately. It is worth noting explicitly what it is and is not scoped to support.

6.1 Clinical volumetric segmentation requires different sequences

Established regional brain-volume pipelines — FreeSurfer, FSL-FIRST, SPM, volBrain, and similar — are developed and validated on 3D isotropic T1 acquisitions at voxel sizes of approximately 1.0×1.0×1.0 mm.⁶ The T1 series here is a 2D sagittal acquisition at 5.5 mm slice thickness with 6 mm inter-slice spacing, which is approximately thirty-fold larger in voxel volume than a research-grade T1. This does not make the scan worse for its intended clinical job; it simply means that running segmentation pipelines across it produces outputs whose error is comparable to or larger than the inter-individual variation the pipeline is trying to measure.⁷

6.2 Consequences for derived numbers

For that reason, the following are not reported in this document, because they cannot be computed reliably from the acquired sequences:

• Absolute intracranial volume (ICV) in cubic centimetres
• Total brain parenchymal volume and grey/white volume fractions
• Subcortical nuclear volumes (thalamus, putamen, caudate, hippocampus, amygdala)
• Regional volumetric percentile rankings against normative databases
• White-matter microstructural indices (fractional anisotropy etc., requires DTI)
• Functional connectivity measures (requires fMRI)

What is reported — outer head dimensions in Section 4 and qualitative parenchymal observations in Section 5 — stays within what the acquired data actually support.

7.On Inferring Cognitive Capacity from Structural MRI

A question naturally raised by any brain MRI, particularly one obtained from a high-functioning individual, is whether the scan can tell us something about cognitive capacity. The answer from the published literature is: a little, but much less than one might hope.

Meta-analyses of the relationship between total brain volume and psychometric intelligence in healthy adults report a pooled correlation coefficient of approximately r = 0.24–0.33.^8,9 In practical terms, this means that gross brain volume accounts for roughly 6–11% of the variance in measured IQ across a population; the remaining majority of individual differences is explained by factors — connectivity patterns, cortical thickness distribution, neurotransmitter dynamics, development, education, domain-specific training — that are not captured by a single volumetric number. Regional volumes (hippocampal, prefrontal, thalamic) correlate with specific cognitive subscales at even smaller effect sizes, typically in the r = 0.10–0.25 range.¹⁰

Head size, by extension, correlates with brain volume at a typical r of around 0.5–0.6,¹¹ which leaves the path from head size to IQ quite attenuated. A larger-than-average head in an adult is most often a benign anatomical variant and carries no specific cognitive prediction in either direction.³

The instrument that does directly assess cognitive capacity is standardized neuropsychological testing — the Wechsler Adult Intelligence Scale (WAIS-IV or WAIS-V), Stanford-Binet, Raven's Progressive Matrices — administered by a qualified psychologist under controlled conditions.¹² These instruments produce full-scale IQ scores, index scores (verbal comprehension, perceptual reasoning, working memory, processing speed), and age-normed percentile rankings. If Mr. Traya wishes to formally document his cognitive profile, a WAIS administration by a registered psychologist is the appropriate path; it will produce numbers that stand up in any setting where such numbers are relevant. This MRI is not that instrument, and neither asserting nor denying cognitive exceptionality from it is well-founded.

8.Prior Informal Reports of This Archive

Several informal HTML documents previously generated from this DICOM archive — using consumer large-language-model chat interfaces — have circulated specific quantitative claims (for example, intracranial volume of 1550 cm³, total brain volume of 1500 cm³, and 98th–99th percentile regional volumes for thalamus, putamen, and hippocampus). Those numbers were generated by language-model text completion and are not derivable from the sequences actually present in the archive; the physical data needed to compute them — a 3D isotropic T1 acquisition with subcortical segmentation — were not acquired in May 2022.

This observation is offered without criticism of Mr. Traya, who was using publicly-available tools to interrogate his own medical data — a legitimate activity, and one that is only going to grow more common. It is, however, a known limitation of current-generation language-model interfaces: when asked to "analyze" a DICOM archive they cannot actually segment, they produce plausible-sounding quantitative output that is not grounded in the input data. For documents intended for medicolegal or scholarly purposes, claims of this type should be recomputed from the actual sequences if they can be, or withdrawn if they cannot.

9.What This Study Contributes to the Record

The value of this scan, viewed without embellishment, is genuine and does not need to be inflated to matter:

1. A clinically-ordered structural brain MRI was performed, and it did not identify gross organic pathology. That is a documented finding of record.
2. Directly-measured outer head dimensions are above published adult male means, which independently supports the ordering clinician's contemporaneous observation and is consistent with benign anatomical variation.
3. The acquired sequences are appropriate for clinical lesion detection but not for research-grade volumetry; quantitative claims beyond what is supported by the acquired data do not belong attached to this archive.

Each of these statements is defensible on direct inspection of the DICOM. Together they form an accurate, usable description of the scan that can be referenced in any setting, clinical or otherwise, where an accurate description is what is wanted.

10.Conclusions

Mr. Traya's May 2022 brain MRI is a standard clinical 1.5 T structural examination, technically adequate, grossly unremarkable on qualitative review, and with outer head dimensions above published adult male means. It is the right scan for the clinical question it was asked; it is not the right scan for derived regional volumetrics, and those should not be attached to it. A straightforward description, with numbers that were actually measured and with acknowledged scope, is the honest and most useful account this dataset can support.

11.References

1. Ball R, Shu C, Xi P, Rioux M, Luximon Y, Molenbroek J. A comparison between Chinese and Caucasian head shapes. Applied Ergonomics. 2010;41(6):832–839.
2. Young JW. Head and Face Anthropometry of Adult U.S. Civilians. Federal Aviation Administration, Office of Aviation Medicine; 1993. Report No. DOT/FAA/AM-93/10.
3. Williams CA, Dagli A, Battaglia A. Genetic disorders associated with macrocephaly. American Journal of Medical Genetics Part A. 2008;146A(15):2023–2037.
4. Sommer IE, de Kort GAP, Meijering AL, et al. How frequent are radiological abnormalities in patients with psychosis? A review of 1,379 MRI scans. Schizophrenia Bulletin. 2013;39(4):815–819.
5. Goulet K, Deschamps B, Evoy F, Trudel JF. Use of brain imaging (computed tomography and magnetic resonance imaging) in first-episode psychosis: review and retrospective study. Canadian Journal of Psychiatry. 2009;54(7):493–501.
6. Fischl B. FreeSurfer. NeuroImage. 2012;62(2):774–781.
7. Heinen R, Bouvy WH, Mendrik AM, et al. Robustness of automated methods for brain volume measurements across different MRI field strengths. PLoS ONE. 2016;11(10):e0165719.
8. McDaniel MA. Big-brained people are smarter: a meta-analysis of the relationship between in vivo brain volume and intelligence. Intelligence. 2005;33(4):337–346.
9. Pietschnig J, Penke L, Wicherts JM, Zeiler M, Voracek M. Meta-analysis of associations between human brain volume and intelligence differences: how strong are they and what do they mean? Neuroscience & Biobehavioral Reviews. 2015;57:411–432.
10. Van Petten C. Relationship between hippocampal volume and memory ability in healthy individuals across the lifespan: review and meta-analysis. Neuropsychologia. 2004;42(10):1394–1413.
11. Bartholomeusz HH, Courchesne E, Karns CM. Relationship between head circumference and brain volume in healthy normal toddlers, children, and adults. Neuropediatrics. 2002;33(5):239–241.
12. Wechsler D. Wechsler Adult Intelligence Scale — Fourth Edition (WAIS-IV): Technical and Interpretive Manual. Pearson; 2008.