Table 4.

Recurring common properties of DTI profiles, arranged by expected order of white matter reversibility.

Order	Occurrence	DTI profile properties	White matter injury patterns	Key references supporting interpretation of injury
I.	Difference/ change	No significant difference/ change in cohorts vs. controls or themselves.	Preserved integrity	For diffusion patterns affected by aging, (26, 27).
II.	Difference/ change	Small change in FA, concurrent and significant decreases in MD, L1 and L2 and 3. With further recovery, increase in FA due to significant increase in L1, decrease in L2 and 3.	Consistent with a range of processes implied by the mechanisms of Neural Repair	Dynamic changes in DTI profiles in subacute injury appear as DTI conflicts (see Order VIII; at-risk). Recovery is seen with FA increase, significant L1 increase and L2 and 3 decrease (20, 28). In chronic TBI, FA increase, correlated with cognitive functioning, is suggestive of neuroplasticity (29, 30).
III.	Change only	Decrease in L1 but with $\frac{1}{2}$>2.5-fold (or much higher) increase in L2 and 3. A decrease in MD is typical. Decrease in FA due to previous increase in compression, proportional to improvement in L1/L2 and 3.	Improvement in Compression	The hallmark of compression, a predominant increase in L1, is remediable with intervention across hydrocephalic conditions, from acute to chronic (9, 31–33). For this contradictory profile in early and late post-operative stages, Keong et al. (10) and this study. For MD decrease, Ivkovic et al. (34). For a decrease in post-operative FA only in shunt responders, Kanno et al. (19).
IV.	Difference	Driven by significant increase in L1, disproportionate >2.0-fold vs. MD/L2and3 measures. Increase in FA.	Compression	Typical of acute pediatric hydrocephalus, (31, 32). For NPH, in remote functional white matter families, e.g., PLIC, Hattori et al. (21) and this study. For FA and L1 increases in pediatric hydrocephalus, (35).
V.	Difference	Increase in L1 predominant (or predominant decrease in L2 and 3), significant increase in FA/MD.	Stretch/ Compression	For “predominant stretch/compression” in NPH/ hydrocephalus, (9, 10, 21, 31, 36). In brain tumors, displacement causes fiber tension or high alignment, Schonberg et al. (15).
VI.	Difference/ change	Driven by L2and3, highly disproportionate increases >2.5-fold vs. MD/L1 measures. FA may be decreased, even significant	Distortion predominantly due to fluid and /or post-operative hydrocephalus	For “predominant transependymal diffusion with the presence of stretch/compression” in pre-and post-operative NPH, Keong et al. (10). Increased L2 and 3 and MD reflect axonal disruption, reversal of CSF flow through ependyma and expansion of extracellular space (interstitial oedema); increased L1 is due to stretch. For all 3 changes and decreased FA in pediatric hydrocephalus, Mangano et al. (35). For post-operative young adult hydrocephalus, small ventricles; decreased FA driven by L2 and 3, Tan et al. (37). DTI profile in common for post-operative hydrocephalus; seen across subtypes and white matter tracts.
VII.	Difference	Disproportionate differences in L 2and 3 >1.5 to <2.5-fold vs. lowest non-FA measure. Global DTI profile of worsening = concurrent decrease in FA, increases in MD, L1, L2 and 3. If L2and3 increase predominant, follow Order VI/VII. If L2and3 increase <1.5-fold or MD/ FA predominant, follow Order VIII/IX.	Oedema and/or loss of integrity	For vasogenic oedema post-TBI, Mac Donald et al. (38) and Veeramuthu et al. (24). For significant decreases in FA due to increases in L2 and 3 in compressive pituitary tumor patients with demyelination, and DTI variability along optic tracts due to anatomy, Paul et al. (39). LI followed by MD increase, were the most sensitive markers in Mild Cognitive Impairment; in MCI and Alzheimer's disease (AD), FA was the least sensitive (25). In early prion disease, Lee et al. (40).
Order	Occurrence	DTI profile properties	White matter injury patterns	Key references supporting interpretation of injury
VIII.	Difference/ change	Global DTI profile of worsening; L2 and 3 increase not disproportionate (<1.5-fold) or predominant/significant individual DTI measure of global change (i.e., FA/MD). For Differences -MD increase predominant orFA decrease predominant or FA decrease significant and Global DTI profile; but fails to match Order VI/VII (not disproportionate increases) or FA decrease significant and At-risk profile; decrease in L1, increase in L2 and 3. For changes –MD increase significant or predominant and global DTI profile; but fails to match order VI/ IX (not disproportionate or too few significant increases). At-risk profile; decrease in L1, increase in L2 and 3.	White matter at-risk of injury disruption due to compression/ stretch/oedemaand/or loss of structure/ atrophy	DTI profile of risk of white matter injury across multiple pathologies. For an NPH model of White Matter At-Risk, Keong et al. (10). For oedema and distension in NPH, (41–45). Stretch component if L1 increase significant. For “stretch/oedema,” impact of proximity to ventricle disrupts both axons and periventricular vasculature, causing impaired autoregulation and increased interstitial fluid (10). For a rat model of hydrocephalus, Yuan et al. (46). For ventricular risk/ FA conflicts, (10, 18, 31, 47). For increased MD in pediatric hydrocephalus, reversing with surgery, Isaacs et al. (48). For high self-corrected ΔADC in NPH vs. controls/atrophy, Takatsuji-Nagaso et al. (49). For TBI, Castano-Leon et al. (28) and Sidaros et al. (11); L1 decrease as marker of TBI severity, Lawrence et al. (50). MD increase has been found to be consistently more sensitive than FA decrease in both MCI and AD, across early to late disease (25, 51).
				For Complex vs. Classic NPH; Lock et al. (9). For NPH, MCI and AD; Horinek et al. (52) and Lee et al. (53). For brain tumors, Yuan et al. (54); low anisotropy (q), high isotropy (p), Price et al. (55). For neurodegeneration, (25, 56–64). For decreased FA post-concussion, (12, 30, 65). For vegetative state in ischaemic hypoxic brain injury and TBI, Newcombe et al. (66).
IX.	Change only	FA decrease significant & predominant or3 significant concurrent increases in MD, L1 and L2 and 3. Diffusion in all directions increased with significant loss of microstructural integrity.	Neuronal degeneration	For neurodegeneration, (13, 14, 67–72). For correlation to multimodal MR imaging in atypical AD; Sintini et al. (73). For significant FA reductions, increased with disease duration in prion disease, Lee et al. (40). For axonal degeneration/ demyelination in TBI, Mac Donald et al. (38), Lawrence et al. (50); increases in MD and L2and3. For increases in all 3 non-FA measures in MCI/ AD, Mayo et al. (74) and Bigham et al. (75). In comparing both MCI and AD to controls, effect sizes for MD, L2and3 and L1 were greater than FA; L1 and L2 and 3 increases were the most discriminatory in early vs. late changes respectively (25), Both associated with white matter deficits and clinical impairment; L1 and MD increases being “state-specific,” remaining relatively static with advancing disease, whereas L2and3 increase with FA decrease was “stage-specific,” being increasingly abnormal with disease progression. Longitudinal analysis showed progressive changes in the latter always occurred in areas that had first shown the former; thought to suggest that L1 increase represents an upstream event preceding neuronal loss (51). Despite known limitations of DTI and its variability across scanning sites, consistent findings of early L1 increase and sensitivity of MD over FA as a biomarker of disease reported across AD studies (76). Neurodegenerative changes provide support for DTI profile properties in Orders VII-IX, representing a proposed pattern of progression toward more irreversible injury in the periodic table.
X.	Difference/ change	Significant decrease in FA, driven by significant decreases in MD and L1. L2 and 3 may be either decreased/ increased/equivocal.	Swelling/hyper-acute/acute &/or irreversible injury	For severe TBI patients, Veenith et al. (77) and Lawrence et al. (50). For worse outcome from chronic TBI, Castano-Leon et al. (28). For reactive astrocytic gliosis without neuronal degeneration in prion disease, Caverzasi et al. (78).

Mimicking the concept of periods, “Orders” reflect commonly recurring patterns of DTI profiles in white matter (i.e., their “neural” properties) seen in response to injury. Here, we arrange them from Order I to X in our predicted trend from reversible to irreversible brain injury (see Discussion for published literature); interpretation for white matter injury patterns in italics are the authors' proposed hypotheses. In “Occurrence,” we specify whether these “neural” properties are to be found when examining a Difference between cohorts (Patient Cohorts vs. Controls), a Change between them (Cohorts vs. Themselves across different timepoints) or both. To solve the Order of mapping DTI profiles to the periodic table, we devised a hierarchical algorithm of rules (Table 5). The colour coding for the Periodic table signifies the following interpretation of white matter tissue signatures: Blue - Normal-appearing integrity or Healthy DTI profiles. In the Order spectrum from reversible to irreversible DTI profiles, a code akin to a “traffic light” warning system: Green - Potential for reversible brain injury; Amber - Risk of progressing towards irreversibility of injury; Red - Likelihood of irreversible brain injury.