The effects of FreeSurfer version, workstation type, and Macintosh operating system version on anatomical volume and cortical thickness measurements - PubMed (original) (raw)

The effects of FreeSurfer version, workstation type, and Macintosh operating system version on anatomical volume and cortical thickness measurements

Ed H B M Gronenschild et al. PLoS One. 2012.

Abstract

FreeSurfer is a popular software package to measure cortical thickness and volume of neuroanatomical structures. However, little if any is known about measurement reliability across various data processing conditions. Using a set of 30 anatomical T1-weighted 3T MRI scans, we investigated the effects of data processing variables such as FreeSurfer version (v4.3.1, v4.5.0, and v5.0.0), workstation (Macintosh and Hewlett-Packard), and Macintosh operating system version (OSX 10.5 and OSX 10.6). Significant differences were revealed between FreeSurfer version v5.0.0 and the two earlier versions. These differences were on average 8.8 ± 6.6% (range 1.3-64.0%) (volume) and 2.8 ± 1.3% (1.1-7.7%) (cortical thickness). About a factor two smaller differences were detected between Macintosh and Hewlett-Packard workstations and between OSX 10.5 and OSX 10.6. The observed differences are similar in magnitude as effect sizes reported in accuracy evaluations and neurodegenerative studies.The main conclusion is that in the context of an ongoing study, users are discouraged to update to a new major release of either FreeSurfer or operating system or to switch to a different type of workstation without repeating the analysis; results thus give a quantitative support to successive recommendations stated by FreeSurfer developers over the years. Moreover, in view of the large and significant cross-version differences, it is concluded that formal assessment of the accuracy of FreeSurfer is desirable.

PubMed Disclaimer

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1

Figure 1. Overview of the statistical significance of voxel volume comparisons for all considered structures.

Each cell is color-coded according to the value of −log10(p), ranging from black (_p_>0.05) to white (_p_≤0.00001), see Figure 2 for the color coding scale. The first three columns show the results obtained by comparing HP with Mac workstation for FreeSurfer versions v4.31, v4.5.0, and v5.0.0, respectively. The p values for the differences between the versions v4.3.1, v4.5.0 and v5.0.0 are shown in columns 4 to 6 for the Mac and in columns 7 to 9 for the HP, respectively. Finally, the last three columns refer to the contrast between OSX 10.6 and OSX 10.5 for the three considered FreeSurfer versions. Cells with a small black rectangle inside denote differences which are not significant anymore after FDR correction for multiple comparisons. White cells with an “X” represent structures for which no comparison could be made, such as left and right cerebral cortex and left and right cerebral white matter, because these are no longer available in FreeSurfer v5.0.0. In the heading row, the labels 431, 450, and 500 denote FreeSurfer v4.3.1, v4.5.0, and v5.0.0, respectively.

Figure 2

Figure 2. The same as Figure 1 , but now for the cortical thickness comparisons.

The color coding is represented in 6 categories of −log10(p), in short lnp: black for lnp<1.301 (_p_>0.05); red for lnp<2 (_p_>0.01); orange for lnp<3 (_p_>0.001); gold for lnp<4 (_p_>0.0001); yellow for lnp<5 (_p_>0.00001); white for lnp≥5 (_p_≤0.00001).

Figure 3

Figure 3. The differences in cortical grey matter volumes between FreeSurfer version v4.3.1 and v5.0.0 on a Mac (OSX 10.5).

The upper row shows the left and right percentage absolute volume differences overlaid on the inflated respective hemispheres in lateral and medial views of an average brain (“fsaverage”). The differences are color coded between 0% and 15%, the full range was 2.1% (left precentral gyrus) - 24.9% (right rostral anterior cingulate cortex). The bottom row depicts the corresponding p values (expressed as −log10(p)) of the applied Student t test. The p values are color coded between the FDR level of 1.607 (p = 0.025) and 5.000 (p = 0.00001). The dark grey regions represent sulcal folds and the light grey regions represent gyral folds.

Figure 4

Figure 4. The same as Figure 3 , but now for cortical white matter.

The full range of the difference was 2.0% (left precentral gyrus) - 33.5% (right rostral anterior cingulate cortex).

Figure 5

Figure 5. The differences in subcortical grey matter volumes between FreeSurfer version v4.3.1 and v5.0.0 on a Mac (OSX 10.5).

The upper row shows the percentage absolute volume differences overlaid each time on two typical slices in coronal, saggital, and transversal views of a T1 scan of a subject, transformed to the MNI305 standard space. The differences are color coded between 0% and 15%, the full range was 2.3% (right lateral ventricle) - 59.5% (5th ventricle). The bottom row depicts the corresponding p values (expressed as −log10(p)) of the applied Student t test. The p values are color coded between the FDR level of 1.607 (p = 0.025) and 5.000 (p = 0.00001).

Figure 6

Figure 6. The same as Figure 3 , but now for cortical thickness.

The percentage absolute thickness differences is color coded between 0% and 7%, the full range was 1.2% (right supramarginal gyrus) - 7.7% (right isthmus cingulate cortex). A p value above 0.025 (−log10(p) = 1.602) was statistically significant after applying a FDR correction for multiple comparisons.

Figure 7

Figure 7. Effects of data processing conditions on the voxel volumes for a subsample of (sub)cortical structures.

Panel A shows the detected percentage absolute differences between the results derived on a Mac and HP workstation for three different versions of FreeSurfer. Panel B depicts the differences between FreeSurfer v4.3.1 vs. v4.5.0, v4.3.1 vs. v5.0.0, and v4.5.0 vs. v5.0.0 for the Mac (OSX 10.5) (for HP these are very similar). Panel C displays the differences between OSX 10.6 and OSX 10.5. For comparison purposes the same vertical scale was used as in Figure 3 of Morey et al. in which the same structures up to the left and right thalamus were considered. The significance is indicated at two levels: * : p<0.025 (the FDR level); ** : _p_≤0.0001. Abbreviations: L: left; R: right; Accu: accumbens; Amyg: amygdala; BrStem: brain stem; Caud: caudate; Hipp: hippocampus; LV: lateral ventricle; Pall: pallidum; Puta: putamen; Thal: thalamus; Ento: entorhinal cortex; Fusi: fusiform; Para: parahippocampal gyrus; BrMask: brain mask; Ventr: left+right lateral and inferior lateral ventricles; MITG: medial-inferior temporal gyrus; STG: superior temporal gyrus; TempL: temporal lobe.

Similar articles

Cited by

References

    1. Fischl B. FreeSurfer. Neuroimage. 2012 doi: 10.1016/j.neuroimage.2012.01.021. - DOI - PMC - PubMed
    1. Desikan RS, Segonne F, Fischl B, Quinn BT, Dickerson BC, et al. An automated labeling system for subdividing the human cerebral cortex on MRI scans into gyral based regions of interest. Neuroimage. 2006;31:968–980. - PubMed
    1. Tae WS, Kim SS, Lee KU, Nam EC, Kim KW. Validation of hippocampal volumes measured using a manual method and two automated methods (FreeSurfer and IBASPM) in chronic major depressive disorder. Neuroradiology. 2008;50:569–581. - PubMed
    1. Morey RA, Petty CM, Xu Y, Hayes JP, Wagner HR, 2nd, et al. A comparison of automated segmentation and manual tracing for quantifying hippocampal and amygdala volumes. Neuroimage. 2009;45:855–866. - PMC - PubMed
    1. Lehmann M, Douiri A, Kim LG, Modat M, Chan D, et al. Atrophy patterns in Alzheimer's disease and semantic dementia: a comparison of FreeSurfer and manual volumetric measurements. Neuroimage. 2010;49:2264–2274. - PubMed

Publication types

MeSH terms

LinkOut - more resources