Dense Sampling of Single Participant (Participant01)

Data Snapshot

DaysSinceTBI WeeksSinceTBI_label WeeksSinceTBI Tract FA RD MD fiber_ratio_map fiber1_ratio_map fiber2_ratio_map fiber3_ratio_map cell_ratio_map water_ratio_map perfusion_ratio_map hind_ratio_map Dep_BDI Anx_BAI GNG_MeanTrimmedCorrectGoRT GNG_GoTrialAccuracy GNG_MeanTrimmedNoGoRT GNG_NoGoTrialAccuracy dprimeScore NUMSYM_MeanTrimmedCorrectRT NUMSYM_Accuracy ANTISACC_MeanTrimmedCorrectRT ANTISACC_Accuracy MotionType HeadMotion
1 Week1 1 forceps major 0.5825151 0.000833279 0.000286841 0.000468987 0.8912177 0.7120187 0.09056854 0.08863013 0.01885955 0.05280789 0.01808319 0.01871481 38 24 297.6661 0.985 243.3222 0.48 2.033184 1324.360 0.9310345 906.5447 0.8928571 Absolute 1.15
1 Week1 1 forceps major 0.5825151 0.000833279 0.000286841 0.000468987 0.8912177 0.7120187 0.09056854 0.08863013 0.01885955 0.05280789 0.01808319 0.01871481 38 24 297.6661 0.985 243.3222 0.48 2.033184 1324.360 0.9310345 906.5447 0.8928571 Relative 0.61
1 Week1 1 forceps major 0.5825151 0.000833279 0.000286841 0.000468987 0.8912177 0.7120187 0.09056854 0.08863013 0.01885955 0.05280789 0.01808319 0.01871481 38 24 297.6661 0.985 243.3222 0.48 2.033184 1324.360 0.9310345 906.5447 0.8928571 Combined 0.88
7 Week2 2 forceps major 0.5885705 0.000835230 0.000282993 0.000467072 0.8925706 0.7293279 0.08434395 0.07889841 0.01708641 0.05133848 0.02133046 0.01735954 23 14 267.6253 1.000 263.1091 0.52 2.649414 1419.198 0.9482759 819.2654 0.9285714 Absolute 0.72
7 Week2 2 forceps major 0.5885705 0.000835230 0.000282993 0.000467072 0.8925706 0.7293279 0.08434395 0.07889841 0.01708641 0.05133848 0.02133046 0.01735954 23 14 267.6253 1.000 263.1091 0.52 2.649414 1419.198 0.9482759 819.2654 0.9285714 Relative 0.37
7 Week2 2 forceps major 0.5885705 0.000835230 0.000282993 0.000467072 0.8925706 0.7293279 0.08434395 0.07889841 0.01708641 0.05133848 0.02133046 0.01735954 23 14 267.6253 1.000 263.1091 0.52 2.649414 1419.198 0.9482759 819.2654 0.9285714 Combined 0.54

Head Motion

Relative head motion shows the least variance and will therefore be used for all further analyses. Wherever the variable HeadMotion is used/mentioned, it implies relative head motion.

FA Trajectory

  • FA is Fractional Anistropy.

  • Fractional Anistropy (FA) is a measure of white matter integrity. FA is used to quantify the degree of direction of water diffusion in white matter. Water tends to diffuse more easily along the axonal fibers of white matter tracts than across them.

  • FA values range from 0 to 1, where:

    • 0 indicates isotropic diffusion, meaning water diffuses equally in all directions (e.g., in cerebrospinal fluid or gray matter).

    • 1 indicates completely anisotropic diffusion, meaning water diffuses only along one direction, which is characteristic of healthy, coherent white matter tracts.

  • FA is an important biomarker for assessing white matter integrity following TBI. A decrease in FA has commonly been interpreted as a loss of white matter (WM) integrity. Hypotheses for this loss of WM integrity include axonal injury, demyelination, and/or disruption of microstructural organization.

Model0

FA ~ weeksSinceTBI + HeadMotion

There is no relationship between white matter integrity and weeks since injury when tract information is not accounted for.

Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.5161969 0.0080591 64.0514126 0.0000000
WeeksSinceTBI -0.0002888 0.0002563 -1.1268079 0.2603832
HeadMotion -0.0131557 0.0134393 -0.9788988 0.3281200

Model1_linear

FA ~ Tract + WeeksSinceTBI + Tract x WeeksSinceTBI + HeadMotion

Excerpt from output shown, with emphasis on interaction effects

Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.0718044 0.0294901 2.434864 0.0152868
Centroid_FA 1.0000000 0.0570639 17.524200 0.0000000
Tractforceps minor -0.0694128 0.0018557 -37.404685 0.0000000
Tractright cab:WeeksSinceTBI 0.0002277 0.0001158 1.965548 0.0499681
Tractright cing:WeeksSinceTBI -0.0005019 0.0001158 -4.333233 0.0000182
Tractleft cst:WeeksSinceTBI 0.0002341 0.0001158 2.020743 0.0439008
Tractleft ilf:WeeksSinceTBI 0.0004087 0.0001158 3.528508 0.0004611
Tractleft slfp:WeeksSinceTBI 0.0003019 0.0001158 2.606657 0.0094479
Tractleft slft:WeeksSinceTBI 0.0002673 0.0001158 2.307521 0.0214807
Tractleft unc:WeeksSinceTBI 0.0003761 0.0001158 3.246574 0.0012557

Interpretation:

  • Intercept: The intercept of 0.0741045 represents the estimated globally averaged FA (Centroid_FA) across all tracts at baseline (i.e., when WeeksSinceTBI = 0 and HeadMotion = 0).
  • Centroid_FA: The estimate of 1.0000000 with a very high t-value indicates that the FA values for individual tracts align very closely with the overall (global) FA trend, suggesting that changes in FA across tracts follow a common trajectory.
  • Tractforceps minor: After accounting for the overall FA pattern across all tracts (represented by Centroid_FA), the FA in forceps minor is specifically 0.06941 units lower than in the reference tract, forceps major.
  • Tractright cab:WeeksSinceTBI: Compared to forceps major, the rate of change in FA over time (WeeksSinceTBI) in the right cab is 0.0002277 units higher for each additional week post-TBI. That is, the right cab experiences a slightly faster increase (or slower decline) in FA over time relative to forceps major, after accounting for the general FA trend across all tracts (as captured by Centroid_FA).

Model2_quadratic

FA ~ Tract + WeeksSinceTBI + WeeksSinceTBI^2 + Tract x WeeksSinceTBI + Tract x WeeksSinceTBI^2 + HeadMotion

Excerpt from output shown, with emphasis on interaction effects

Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.0741045 0.0340678 2.175209 0.0301586
Centroid_FA 1.0000000 0.0655240 15.261574 0.0000000
Tractforceps minor -0.0710114 0.0028369 -25.030920 0.0000000
Tractleft unc -0.1214744 0.0028369 -42.818711 0.0000000
Tractright unc -0.1236779 0.0028369 -43.595423 0.0000000
Tractleft atr:WeeksSinceTBI 0.0010033 0.0004670 2.148541 0.0322286
Tractleft cab:WeeksSinceTBI 0.0018443 0.0004670 3.949514 0.0000915
Tractright cab:WeeksSinceTBI 0.0019140 0.0004670 4.098859 0.0000496
Tractright unc:WeeksSinceTBI 0.0011533 0.0004670 2.469705 0.0139098
Tractleft cab:I(WeeksSinceTBI^2) -0.0000484 0.0000162 -2.990939 0.0029411
Tractright cab:I(WeeksSinceTBI^2) -0.0000602 0.0000162 -3.720952 0.0002248
Tractright unc:I(WeeksSinceTBI^2) -0.0000419 0.0000162 -2.586419 0.0100255

Interpretation:

Tractleft cab: Early after TBI, FA in the left cingulum bundle shows a small increase per week (0.00184 units per the linear tractleft cab:WeekSinceTBI interaction). As time progresses, the rate of this increase slows down due to the negative quadratic effect (-0.000484 units per squared week) and may lead to a flattening or decline [to baseline] if the deceleration continues long enough.

ModelCHOICE_linearOrQuadratic?

Quadratic model is a better fit for describing the relationship between FA and time since injury across all tracts at alpha=0.05.

Therefore, we consider results from Model2_Quadratic to assess the trajectory of FA across tracts.

model Res.Df RSS Df Sum of Sq F Pr(>F)
linear 448 0.0049229 NA NA NA NA
quadratic 430 0.0044593 18 0.0004635 2.483197 0.0007244

AD Trajectory

  • AD is Axial Diffusivity.

  • Axial diffusivity (AD) represents the diffusion of water molecules along the primary axis of a white matter tract (i.e.,the principal direction of diffusion).

  • Changes in axial integrity may be indicative of axonal damage following TBI.

Model0

AD ~ weeksSinceTBI + HeadMotion

There is no relationship between white matter integrity and weeks since injury when tract information is not accounted for.

Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.0007742 1.15e-05 67.4982048 0.0000000
WeeksSinceTBI 0.0000002 4.00e-07 0.5860182 0.5581370
HeadMotion -0.0000047 1.91e-05 -0.2480542 0.8041979

Model1_linear

AD ~ Tract + WeeksSinceTBI + Tract x WeeksSinceTBI + HeadMotion

Only main effects of time, motion, and interaction effects shown below:

Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.0007361 4.63e-05 15.905973 0.0000000
Centroid_FA 0.0001895 8.95e-05 2.115754 0.0349169
WeeksSinceTBI 0.0000005 1.00e-07 3.947163 0.0000918
Tractright atr:WeeksSinceTBI -0.0000004 2.00e-07 -2.007086 0.0453408
Tractright cing:WeeksSinceTBI -0.0000007 2.00e-07 -4.033525 0.0000646
Tractright cst:WeeksSinceTBI -0.0000004 2.00e-07 -2.128206 0.0338652
Tractright slfp:WeeksSinceTBI -0.0000006 2.00e-07 -3.215181 0.0013979
Tractright slft:WeeksSinceTBI -0.0000004 2.00e-07 -2.400754 0.0167682

Model2_quadratic

AD ~ Tract + WeeksSinceTBI + WeeksSinceTBI^2 + Tract x WeeksSinceTBI + Tract x WeeksSinceTBI^2 + HeadMotion

Only main effects of time, motion, and interaction effects shown below:

Estimate Std. Error t value Pr(>|t|)
WeeksSinceTBI 1.2e-06 5e-07 2.308189 0.0214623
Tractright cing:WeeksSinceTBI -2.7e-06 7e-07 -3.645031 0.0003001
Tractright cing:I(WeeksSinceTBI^2) 1.0e-07 0e+00 2.736064 0.0064754

ModelCHOICE_linearOrQuadratic?

Quadratic model is not a significantly better fit for describing the relationship between AD and time since injury across all tracts (at alpha=0.05).

Therefore, we consider results from Model1_linear to assess the trajectory of AD across tracts.

model Res.Df RSS Df Sum of Sq F Pr(>F)
linear 448 0 NA NA NA NA
quadratic 430 0 18 0 1.910131 0.0137679

RD Trajectory

  • RD is Radial Diffusivity.

  • Radial diffusivity (RD) represents the degree of water diffusion perpendicular to the main direction of white matter fibers (i.e., across axons and their surrounding myelin sheaths).

  • It is used to index changes in the myelin sheath around axons. Increases in RD are often interpreted as indicating myelin damage or demyelination, where the insulating layer around the axons is compromised.

Model0

RD ~ weeksSinceTBI + HeadMotion

There is no relationship between white matter integrity and weeks since injury when tract information is not accounted for.

Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.0003207 6.70e-06 47.6604591 0.0000000
WeeksSinceTBI 0.0000003 2.00e-07 1.2807431 0.2008986
HeadMotion 0.0000072 1.12e-05 0.6437709 0.5200296

Model1_linear

RD ~ Tract + WeeksSinceTBI + Tract x WeeksSinceTBI + HeadMotion

Only main effects of time, motion, and interaction effects shown below:

Estimate Std. Error t value Pr(>|t|)
WeeksSinceTBI 5.0e-07 1e-07 5.675347 0.0000000
HeadMotion 7.2e-06 1e-06 7.311057 0.0000000
Tractleft atr:WeeksSinceTBI -2.0e-07 1e-07 -2.180476 0.0297407
Tractleft cab:WeeksSinceTBI -4.0e-07 1e-07 -3.413750 0.0006989
Tractleft cing:WeeksSinceTBI -2.0e-07 1e-07 -2.016118 0.0443826
Tractleft cst:WeeksSinceTBI -3.0e-07 1e-07 -2.228391 0.0263489
Tractright cst:WeeksSinceTBI -3.0e-07 1e-07 -2.807188 0.0052152
Tractleft ilf:WeeksSinceTBI -3.0e-07 1e-07 -2.471847 0.0138115
Tractleft slfp:WeeksSinceTBI -3.0e-07 1e-07 -2.950630 0.0033372
Tractright slfp:WeeksSinceTBI -3.0e-07 1e-07 -2.663475 0.0080119
Tractleft slft:WeeksSinceTBI -3.0e-07 1e-07 -2.761581 0.0059881
Tractleft unc:WeeksSinceTBI -3.0e-07 1e-07 -2.438989 0.0151148

Model2_quadratic

RD ~ Tract + WeeksSinceTBI + WeeksSinceTBI^2 + Tract x WeeksSinceTBI + Tract x WeeksSinceTBI^2 + HeadMotion

Only main effects of time, motion, and interaction effects shown below:

Estimate Std. Error t value Pr(>|t|)
WeeksSinceTBI 1.6e-06 3e-07 5.104004 0.0000005
I(WeeksSinceTBI^2) 0.0e+00 0e+00 -3.754294 0.0001977
HeadMotion 6.0e-06 9e-07 6.461835 0.0000000
Tractleft atr:WeeksSinceTBI -1.0e-06 4e-07 -2.248009 0.0250804
Tractleft cab:WeeksSinceTBI -1.1e-06 4e-07 -2.564894 0.0106583
Tractright cab:WeeksSinceTBI -1.2e-06 4e-07 -2.730172 0.0065896
Tractleft cing:WeeksSinceTBI -9.0e-07 4e-07 -1.991556 0.0470504
Tractleft cst:WeeksSinceTBI -1.0e-06 4e-07 -2.213939 0.0273557
Tractright cst:WeeksSinceTBI -1.2e-06 4e-07 -2.819264 0.0050353
Tractleft slfp:WeeksSinceTBI -1.1e-06 4e-07 -2.443104 0.0149620
Tractleft slft:WeeksSinceTBI -1.0e-06 4e-07 -2.240735 0.0255518
Tractright unc:WeeksSinceTBI -1.1e-06 4e-07 -2.460135 0.0142799
Tractright cab:I(WeeksSinceTBI^2) 0.0e+00 0e+00 2.291418 0.0224208
Tractright cst:I(WeeksSinceTBI^2) 0.0e+00 0e+00 2.157677 0.0315048
Tractright unc:I(WeeksSinceTBI^2) 0.0e+00 0e+00 2.458337 0.0143506

ModelCHOICE_linearOrQuadratic?

Quadratic model is a significantly better fit for describing the relationship between RD and time since injury across all tracts (at alpha=0.05).

Therefore, we consider results from Model2_Quadratric to assess the trajectory of RD across tracts.

model Res.Df RSS Df Sum of Sq F Pr(>F)
linear 449 0 NA NA NA NA
quadratic 431 0 18 0 4.72133 0

MD Trajectory

  • MD is Mean Diffusivity.

  • Mean diffusivity (MD) represents the average diffusion of water molecules within a voxel. It reflects the overall magnitude of water diffusion in all directions, without favoring any particular direction (unlike FA, AD, and MD).

  • As a non-directional measure of tissue integrity, it’s interpretation is as follows:

    • Low MD: Indicates restricted diffusion, which can be due to dense cellular structures, such as in healthy white matter. However, in pathological conditions like edema or cell swelling, it can increase due to the buildup of water content.

    • High MD: Suggests increased water diffusion, which may reflect tissue degeneration, loss of cellular barriers (such as in atrophy or demyelination), or the presence of free extracellular water.

Model0

MD ~ weeksSinceTBI + HeadMotion

There is no relationship between white matter integrity and weeks since injury when tract information is not accounted for.

Estimate Std. Error t value Pr(>|t|)
(Intercept) 0.0004719 6.90e-06 68.1212415 0.0000000
WeeksSinceTBI 0.0000003 2.00e-07 1.1528500 0.2495422
HeadMotion 0.0000032 1.16e-05 0.2799678 0.7796221

Model1_linear

MD ~ Tract + WeeksSinceTBI + Tract x WeeksSinceTBI + HeadMotion

Only main effects of time, motion, and interaction effects shown below:

Estimate Std. Error t value Pr(>|t|)
WeeksSinceTBI 5.0e-07 1e-07 5.720889 0.0000000
HeadMotion 3.2e-06 1e-06 3.273718 0.0011434
Tractleft atr:WeeksSinceTBI -3.0e-07 1e-07 -2.277714 0.0232129
Tractleft cab:WeeksSinceTBI -3.0e-07 1e-07 -2.684222 0.0075387
Tractleft cing:WeeksSinceTBI -3.0e-07 1e-07 -2.400315 0.0167871
Tractleft cst:WeeksSinceTBI -2.0e-07 1e-07 -2.147678 0.0322739
Tractright cst:WeeksSinceTBI -3.0e-07 1e-07 -3.016475 0.0027025
Tractright ilf:WeeksSinceTBI 2.0e-07 1e-07 2.046971 0.0412425
Tractleft slfp:WeeksSinceTBI -3.0e-07 1e-07 -2.927912 0.0035859
Tractright slfp:WeeksSinceTBI -4.0e-07 1e-07 -3.505193 0.0005021
Tractleft slft:WeeksSinceTBI -3.0e-07 1e-07 -2.844818 0.0046468

Model2_quadratic

MD ~ Tract + WeeksSinceTBI + WeeksSinceTBI^2 + Tract x WeeksSinceTBI + Tract x WeeksSinceTBI^2 + HeadMotion

Only main effects of time, motion, and interaction effects shown below:

Estimate Std. Error t value Pr(>|t|)
WeeksSinceTBI 1.3e-06 3e-07 4.216100 0.0000303
I(WeeksSinceTBI^2) 0.0e+00 0e+00 -2.878079 0.0042000
HeadMotion 2.3e-06 1e-06 2.430938 0.0154666
Tractleft cing:WeeksSinceTBI -1.0e-06 5e-07 -2.308348 0.0214523
Tractright cing:WeeksSinceTBI -1.1e-06 5e-07 -2.359973 0.0187211
Tractleft cst:WeeksSinceTBI -9.0e-07 5e-07 -2.025018 0.0434815
Tractright cst:WeeksSinceTBI -1.2e-06 5e-07 -2.711233 0.0069712
Tractleft slfp:WeeksSinceTBI -1.0e-06 5e-07 -2.157468 0.0315212
Tractright cing:I(WeeksSinceTBI^2) 0.0e+00 0e+00 2.097396 0.0365405
Tractright cst:I(WeeksSinceTBI^2) 0.0e+00 0e+00 2.018275 0.0441816

ModelCHOICE_linearOrQuadratic?

Quadratic model is a significantly better fit for describing the relationship between MD and time since injury across all tracts (at alpha=0.05).

Therefore, we consider results from Model2_Quadratric to assess the trajectory of MD across tracts.

model Res.Df RSS Df Sum of Sq F Pr(>F)
linear 449 0 NA NA NA NA
quadratic 431 0 18 0 2.78334 0.0001375

DiffusionMetric: Choice

AIC determined: RD

Diffusion Metric AIC Linear Model AIC Quadratic Model
FA -3880.108 -3935.517
AD -10405.38 -10399.95
RD -10873.97 -10925.44
MD -10874.13 -10891.58

RD shows the most change by tract

Questions

  • FA has better scaling properties (E.g., range b/w 0.45 to 0.60), versus other metrics which show differences in a range of 0.0003 - 0.00085 (say).

  • Go with FA, but also show variability in other metrics with maximal tract*time interaction?

DiffusionMetric: Plots

Across Tracts

By Tract

Forceps

Major
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.52 0.81
AD 0.50 0.54
RD 0.56 0.81
MD 0.58 0.75

Minor
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.44 0.54
AD 0.08 0.08
RD 0.34 0.41
MD 0.23 0.26

Anterior Thalamic Radiation

Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.1 0.14
AD 0.15 0.19
RD 0.15 0.2
MD 0.18 0.24

Cingulum

Angular Bundle
Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.22 0.22
AD 0.15 0.25
RD 0.35 0.4
MD 0.24 0.36

Right
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.46 0.53
AD 0.17 0.25
RD 0.37 0.37
MD 0.24 0.28

Cingulate Gyrus
Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.1 0.18
AD 0.02 0.06
RD 0.16 0.23
MD 0.15 0.16

Right
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.65 0.72
AD 0.06 0.18
RD 0.74 0.79
MD 0.41 0.41

Cortico-Spinal

Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.24 0.3
AD 0.23 0.24
RD 0.3 0.35
MD 0.33 0.35

Right
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.53 0.58
AD 0.06 0.09
RD 0.57 0.59
MD 0.28 0.28

Inferior Longitudinal Fasciculus

Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.13 0.33
AD 0.46 0.52
RD 0.17 0.47
MD 0.33 0.55

Right
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.6 0.69
AD 0.58 0.63
RD 0.67 0.76
MD 0.67 0.75

Superior Longitudinal Fasciculus

Parietal
Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.28 0.46
AD 0.12 0.12
RD 0.25 0.35
MD 0.19 0.23

Right
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.44 0.70
AD 0.12 0.14
RD 0.31 0.59
MD 0.1 0.25

Temporal
Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.26 0.41
AD 0.19 0.21
RD 0.21 0.37
MD 0.18 0.29

Right
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.59 0.78
AD 0.05 0.11
RD 0.58 0.73
MD 0.48 0.55

Uncinate

Left
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.1 0.20
AD 0.28 0.31
RD 0.19 0.32
MD 0.25 0.35

Right
R² Values for Linear and Quadratic Models
Metric R2_Linear R2_Quadratic
FA 0.46 0.46
AD 0.17 0.18
RD 0.42 0.42
MD 0.35 0.35

Descriptive Table

Maxima/Minima of Quadratic Fitted Model
Tract Metric MinValue MaxValue R_Sqrd_AIC_Linear R_Sqrd_AIC_Quadratic BetaTime_Linear BetaTime_Quadratic BetaTimeSqrd_Quadratic Vertex_Within_Range Vertex_Value ParabolaDirection
forceps major FA 0.5683, Week: 20 0.5886, Week: 2 0.524, AIC: -225.963 0.806, AIC: -248.143 -4.1e-04*** -1.68e-03*** 4.57e-05*** Yes 18.437 /\
forceps major AD 8e-04, Week: 5 8e-04, Week: 27 0.505, AIC: -584.972 0.536, AIC: -584.727 4.5e-07*** 9.89e-07* -1.93e-08 Yes 25.592 \/
forceps major RD 3e-04, Week: 2 3e-04, Week: 20 0.561, AIC: -601.285 0.813, AIC: -622.281 4.43e-07*** 1.65e-06*** -4.31e-08*** Yes 19.099 \/
forceps major MD 5e-04, Week: 5 5e-04, Week: 20 0.582, AIC: -601.992 0.746, AIC: -613.423 4.45e-07*** 1.43e-06*** -3.52e-08*** Yes 20.288 \/
forceps minor FA 0.4999, Week: 18 0.5233, Week: 3 0.444, AIC: -216.252 0.54, AIC: -219.406 -3.08e-04** -1.14e-03** 2.96e-05* Yes 19.143 /\
forceps minor AD 8e-04, Week: 4 8e-04, Week: 21 0.084, AIC: -566.346 0.084, AIC: -564.352 2.2e-07 2.65e-07 -1.61e-09 No 82.386 \/
forceps minor RD 3e-04, Week: 3 3e-04, Week: 18 0.34, AIC: -592.095 0.413, AIC: -593.272 2.77e-07** 9.05e-07* -2.25e-08 Yes 20.121 \/
forceps minor MD 5e-04, Week: 4 5e-04, Week: 21 0.228, AIC: -586.703 0.262, AIC: -585.9 2.58e-07* 6.92e-07 -1.55e-08 Yes 22.274 \/
left atr FA 0.456, Week: 7 0.4798, Week: 3 0.099, AIC: -205.938 0.141, AIC: -205.252 -1.78e-04 -7.02e-04 1.88e-05 Yes 18.705 /\
left atr AD 7e-04, Week: 2 7e-04, Week: 20 0.149, AIC: -584.324 0.19, AIC: -583.671 1.82e-07 6.62e-07 -1.72e-08 Yes 19.263 \/
left atr RD 3e-04, Week: 6 3e-04, Week: 7 0.15, AIC: -591.803 0.201, AIC: -591.456 1.97e-07 6.58e-07 -1.65e-08 Yes 19.900 \/
left atr MD 5e-04, Week: 5 5e-04, Week: 20 0.176, AIC: -596.801 0.236, AIC: -596.858 1.92e-07* 6.6e-07 -1.68e-08 Yes 19.682 \/
right atr FA 0.454, Week: 18 0.4822, Week: 3 0.555, AIC: -209.724 0.703, AIC: -218.634 -4.56e-04*** -1.75e-03*** 4.63e-05** Yes 18.881 /\
right atr AD 7e-04, Week: 10 7e-04, Week: 1 0.027, AIC: -575.631 0.028, AIC: -573.651 1.06e-07 3.63e-08 2.49e-09 No -7.279 /\
right atr RD 3e-04, Week: 3 3e-04, Week: 18 0.523, AIC: -595.932 0.634, AIC: -601.025 3.5e-07*** 1.19e-06** -3.02e-08* Yes 19.745 \/
right atr MD 5e-04, Week: 4 5e-04, Week: 18 0.367, AIC: -595.901 0.427, AIC: -596.575 2.69e-07** 8.08e-07* -1.93e-08 Yes 20.908 \/
left cab FA 0.4504, Week: 7 0.4669, Week: 16 0.219, AIC: -215.344 0.224, AIC: -213.521 7.24e-05 2.36e-04 -5.85e-06 Yes 20.146 \/
left cab AD 8e-04, Week: 5 9e-04, Week: 18 0.154, AIC: -563.93 0.251, AIC: -565.218 3.33e-07* 1.41e-06* -3.85e-08 Yes 18.282 \/
left cab RD 4e-04, Week: 2 4e-04, Week: 7 0.349, AIC: -617.944 0.405, AIC: -618.378 7.29e-08 4.16e-07 -1.23e-08 Yes 16.927 \/
left cab MD 5e-04, Week: 12 5e-04, Week: 18 0.237, AIC: -605.202 0.357, AIC: -607.833 1.6e-07* 7.46e-07* -2.1e-08* Yes 17.754 \/
right cab FA 0.454, Week: 25 0.472, Week: 2 0.462, AIC: -228.683 0.526, AIC: -230.141 -1.95e-04* 3.52e-04 -1.96e-05 Yes 8.989 \/
right cab AD 8e-04, Week: 4 9e-04, Week: 20 0.171, AIC: -563.503 0.252, AIC: -564.31 2.79e-07 1.28e-06 -3.6e-08 Yes 17.831 \/
right cab RD 4e-04, Week: 4 4e-04, Week: 13 0.367, AIC: -605.862 0.372, AIC: -604.069 2.33e-07** 3.61e-07 -4.57e-09 No 39.512 \/
right cab MD 5e-04, Week: 4 6e-04, Week: 20 0.237, AIC: -594.135 0.278, AIC: -593.623 2.49e-07* 6.69e-07 -1.5e-08 Yes 22.218 \/
left cing FA 0.5776, Week: 13 0.604, Week: 6 0.096, AIC: -193.793 0.182, AIC: -194.498 -2.41e-04 -1.17e-03 3.33e-05 Yes 17.578 /\
left cing AD 8e-04, Week: 5 9e-04, Week: 1 0.02, AIC: -563.09 0.064, AIC: -562.318 1.14e-07 -5.66e-07 2.44e-08 Yes 11.624 /\
left cing RD 3e-04, Week: 6 3e-04, Week: 13 0.162, AIC: -589.509 0.232, AIC: -589.85 2.14e-07* 7.83e-07 -2.04e-08 Yes 19.189 \/
left cing MD 5e-04, Week: 6 5e-04, Week: 20 0.149, AIC: -596.006 0.155, AIC: -594.212 1.8e-07 3.33e-07 -5.48e-09 No 30.393 \/
right cing FA 0.5816, Week: 22 0.623, Week: 3 0.649, AIC: -195.286 0.715, AIC: -198.944 -9.3e-04*** -2.2e-03*** 4.56e-05* Yes 24.156 /\
right cing AD 8e-04, Week: 7 9e-04, Week: 1 0.059, AIC: -546.397 0.183, AIC: -548.201 -2.68e-07 -1.86e-06* 5.7e-08 Yes 16.306 /\
right cing RD 3e-04, Week: 3 3e-04, Week: 22 0.739, AIC: -603.261 0.789, AIC: -606.891 6.08e-07*** 1.27e-06*** -2.38e-08* Yes 26.714 \/
right cing MD 5e-04, Week: 4 5e-04, Week: 18 0.407, AIC: -599.602 0.409, AIC: -597.679 3.16e-07*** 2.28e-07 3.14e-09 No -36.328 /\
left cst FA 0.5041, Week: 5 0.5263, Week: 3 0.24, AIC: -203.548 0.304, AIC: -203.903 -1.87e-04 -9.13e-04 2.6e-05 Yes 17.558 /\
left cst AD 7e-04, Week: 5 7e-04, Week: 20 0.234, AIC: -588.489 0.239, AIC: -586.634 2.37e-07* 8.93e-08 5.28e-09 No -8.450 /\
left cst RD 3e-04, Week: 6 3e-04, Week: 13 0.303, AIC: -604.538 0.352, AIC: -604.492 2.02e-07* 5.98e-07 -1.42e-08 Yes 21.097 \/
left cst MD 4e-04, Week: 6 4e-04, Week: 13 0.334, AIC: -614.318 0.354, AIC: -613.126 2.14e-07** 4.28e-07 -7.68e-09 No 27.871 \/
right cst FA 0.5066, Week: 5 0.5274, Week: 3 0.533, AIC: -225.867 0.581, AIC: -226.779 -2.03e-04* -7.34e-04* 1.9e-05 Yes 19.280 /\
right cst AD 7e-04, Week: 4 7e-04, Week: 1 0.06, AIC: -584.769 0.092, AIC: -583.719 7.49e-08 -3.26e-07 1.44e-08 Yes 11.351 /\
right cst RD 3e-04, Week: 2 3e-04, Week: 5 0.575, AIC: -632.073 0.586, AIC: -630.787 1.43e-07** 2.88e-07 -5.2e-09 No 27.721 \/
right cst MD 4e-04, Week: 4 4e-04, Week: 20 0.28, AIC: -624.128 0.281, AIC: -622.162 1.2e-07* 8.34e-08 1.32e-09 No -31.483 /\
left ilf FA 0.5, Week: 7 0.5207, Week: 3 0.128, AIC: -214.907 0.331, AIC: -220.089 -3.17e-06 -9.87e-04* 3.52e-05* Yes 14.001 /\
left ilf AD 8e-04, Week: 5 9e-04, Week: 18 0.456, AIC: -576.552 0.519, AIC: -577.864 5.28e-07*** 1.38e-06* -3.06e-08 Yes 22.585 \/
left ilf RD 3e-04, Week: 3 4e-04, Week: 7 0.167, AIC: -598.155 0.473, AIC: -608.52 1.7e-07 1.19e-06*** -3.66e-08** Yes 16.277 \/
left ilf MD 5e-04, Week: 2 5e-04, Week: 18 0.33, AIC: -598.518 0.553, AIC: -607.447 2.89e-07** 1.25e-06*** -3.46e-08** Yes 18.137 \/
right ilf FA 0.4814, Week: 18 0.5058, Week: 3 0.605, AIC: -219.569 0.687, AIC: -223.878 -5.44e-04*** -1.4e-03*** 3.05e-05* Yes 22.868 /\
right ilf AD 8e-04, Week: 4 8e-04, Week: 20 0.584, AIC: -575 0.633, AIC: -576.412 7.35e-07*** 1.63e-06** -3.19e-08 Yes 25.472 \/
right ilf RD 3e-04, Week: 4 4e-04, Week: 20 0.674, AIC: -592.609 0.763, AIC: -599.287 6.59e-07*** 1.64e-06*** -3.5e-08** Yes 23.357 \/
right ilf MD 5e-04, Week: 4 5e-04, Week: 20 0.67, AIC: -590.386 0.749, AIC: -595.734 6.84e-07*** 1.63e-06*** -3.4e-08* Yes 24.019 \/
left slfp FA 0.4782, Week: 18 0.4941, Week: 2 0.284, AIC: -231.984 0.456, AIC: -237.402 -1.11e-04 -8.38e-04** 2.6e-05* Yes 16.096 /\
left slfp AD 7e-04, Week: 7 7e-04, Week: 18 0.118, AIC: -598.702 0.12, AIC: -596.759 1.43e-07 6.6e-08 2.75e-09 No -12.003 /\
left slfp RD 3e-04, Week: 2 3e-04, Week: 18 0.253, AIC: -620.808 0.35, AIC: -622.568 1.18e-07* 5.17e-07* -1.43e-08 Yes 18.083 \/
left slfp MD 4e-04, Week: 7 4e-04, Week: 18 0.191, AIC: -618.858 0.227, AIC: -618.07 1.26e-07* 3.67e-07 -8.62e-09 Yes 21.282 \/
right slfp FA 0.4839, Week: 13 0.503, Week: 3 0.442, AIC: -218.726 0.7, AIC: -233.458 -3.37e-04** -1.63e-03*** 4.62e-05*** Yes 17.609 /\
right slfp AD 7e-04, Week: 19 7e-04, Week: 1 0.118, AIC: -602.321 0.144, AIC: -601.124 -1.21e-07 -3.88e-07 9.56e-09 Yes 20.275 /\
right slfp RD 3e-04, Week: 3 3e-04, Week: 13 0.313, AIC: -618.357 0.593, AIC: -630.479 1.52e-07* 8.91e-07*** -2.65e-08*** Yes 16.819 \/
right slfp MD 4e-04, Week: 2 4e-04, Week: 10 0.1, AIC: -626.163 0.246, AIC: -628.945 6.08e-08 4.65e-07* -1.45e-08* Yes 16.058 \/
left slft FA 0.5232, Week: 13 0.5349, Week: 2 0.258, AIC: -234.144 0.413, AIC: -238.457 -1.42e-04 -7.93e-04** 2.33e-05* Yes 17.005 /\
left slft AD 7e-04, Week: 5 8e-04, Week: 19 0.188, AIC: -599.955 0.206, AIC: -598.564 1.2e-07 3.63e-07 -8.71e-09 Yes 20.821 \/
left slft RD 3e-04, Week: 2 3e-04, Week: 18 0.215, AIC: -620.227 0.373, AIC: -624.292 1.35e-07* 6.37e-07** -1.8e-08* Yes 17.708 \/
left slft MD 4e-04, Week: 5 5e-04, Week: 18 0.182, AIC: -619.994 0.294, AIC: -621.965 1.3e-07* 5.45e-07* -1.49e-08 Yes 18.316 \/
right slft FA 0.5001, Week: 18 0.5221, Week: 2 0.587, AIC: -217.462 0.779, AIC: -232.296 -5.48e-04*** -1.87e-03*** 4.74e-05*** Yes 19.746 /\
right slft AD 7e-04, Week: 7 7e-04, Week: 1 0.054, AIC: -605.155 0.112, AIC: -604.865 2.73e-08 -3.39e-07 1.31e-08 Yes 12.914 /\
right slft RD 3e-04, Week: 2 3e-04, Week: 20 0.58, AIC: -614.011 0.73, AIC: -623.922 3.58e-07*** 1.11e-06*** -2.69e-08** Yes 20.623 \/
right slft MD 4e-04, Week: 4 4e-04, Week: 20 0.484, AIC: -624.26 0.553, AIC: -626.093 2.48e-07*** 6.26e-07** -1.35e-08 Yes 23.115 \/
left unc FA 0.4494, Week: 7 0.4727, Week: 3 0.105, AIC: -209.047 0.197, AIC: -209.968 -3.57e-05 -7.62e-04 2.6e-05 Yes 14.642 /\
left unc AD 8e-04, Week: 2 8e-04, Week: 17 0.281, AIC: -575.687 0.312, AIC: -574.844 3.91e-07** 9.14e-07 -1.87e-08 Yes 24.399 \/
left unc RD 4e-04, Week: 3 4e-04, Week: 7 0.186, AIC: -593.985 0.316, AIC: -596.706 1.77e-07 9.06e-07* -2.61e-08* Yes 17.350 \/
left unc MD 5e-04, Week: 2 5e-04, Week: 17 0.251, AIC: -594.642 0.352, AIC: -596.555 2.49e-07* 9.09e-07* -2.36e-08 Yes 19.211 \/
right unc FA 0.4481, Week: 18 0.4706, Week: 12 0.46, AIC: -208.207 0.46, AIC: -206.209 -4.41e-04*** -4.19e-04 -7.77e-07 No -270.041 \/
right unc AD 8e-04, Week: 4 8e-04, Week: 20 0.168, AIC: -567.987 0.177, AIC: -566.281 2.97e-07 6.04e-07 -1.1e-08 No 27.491 \/
right unc RD 4e-04, Week: 4 4e-04, Week: 18 0.423, AIC: -585.384 0.423, AIC: -583.393 4.24e-07*** 4.63e-07 -1.4e-09 No 165.397 \/
right unc MD 5e-04, Week: 4 5e-04, Week: 20 0.35, AIC: -585.446 0.352, AIC: -583.544 3.82e-07** 5.1e-07 -4.59e-09 No 55.504 \/

Model Interpretation

Interpretation of Findings

  • The overall model comparison using ANOVA between the linear (Model1) and quadratic (Model2) models suggests that one model fits the data better globally. The quadratic model explained the data better for FA, RD, and MD overall, while AD was better explained by the linear model. However, this doesn’t necessarily mean that each specific tract will follow the same pattern. For example:

    • If the quadratic model globally fit FA better, it suggests that for the majority of tracts, adding the quadratic term for time (i.e., time^2) improves the model’s ability to capture the variability in the data. But for any given tract (e.g., left cingulum), it is possible that the quadratic term might not be significant or necessary.

    • In instances where only the linear tract*time interaction was significant for a specific tract (when the overall fit with a quadratic model was better), it suggests that the relationship between the diffusion metric and time is best captured linearly for that tract.

Notes: Global fits give you a higher-level understanding of the nature of injury progression (like TBI), how it affects the brain’s white matter integrity, and whether non-linear trajectories are common across different brain regions. This provides broader insights into recovery, repair, or deterioration, which could help design interventions that target the brain system-wide.

Model selection (Fine-Tuning)

Algorithm/Approach

  1. Start with the globally better-fitting model (quadratic or linear) and apply it to individual tracts that came out significant in the tract x time interaction. This ensures that the focus is on tracts which show a meaningful differential change over time across tracts, rather than applying the model indiscriminately.
  2. For each tract, evaluate the significance of the quadratic term for FA and RD tracts (FA and RD only for contrast in min, max values, parabola characteristics):
    • If significant: Retain the quadratic model.
    • If not significant: Consider simplifying the model to linear for that specific tract.
    • Also consider tracts with high R2 values.

Tracts to Focus on

Tracts are selected based on significant tract*time interactions in the globally better-fitting model.

FA:Quadratic

Significant Tracts
left cab
right cab
right unc

AD:Linear

Significant Tracts
right atr
right cing
right cst
right slfp
right slft

RD:Quadratic

Significant Tracts
left atr
left cab
right cab
left cing
left cst
right cst
left slfp
left slft
right unc

MD:Quadratic

Significant Tracts
left cing
right cing
left cst
right cst
left slfp

Manuscript Prep Questions

  • Are there observable patterns in the data in the diffusion metrics, when inspecting by tract? What is the general trend?

  • I think I will need to compare these results to TBIPFM17, and maybe two more TBIPFM participants, to draw a cohesive narrative.

  • What interpretation might we draw in each tract analysis? Can we create a narrative?

  • How are these metrics related to cognition? How are they related to emotion?

  • What are fiber ratio maps? How might they relate to AD, FA, MD, and RD?