Structural cerebellar abnormalities associated with developmental stuttering Beatrice Benedetti1,2,

1
Structural cerebellar abnormalities associated
with developmental stuttering Beatrice
Benedetti1,2, Steve Davis3, Pete Howell3 and Kate
Watkins1,4 1FMRIB Centre, University of Oxford,
UK, 2Neuroimaging Research Unit, Dept. of
Neurology, Ospedale S. Raffaele Milan, Italy,
3Speech Research Group, Dept. Psychology,
University College London, UK, 4Dept.
Experimental Psychology, University of Oxford, UK.
Introduction and Purpose
Results
Developmental stuttering is a disorder affecting
the fluency of speech, characterized by
repetitions, prolongations or difficulties with
initiation of sounds. It affects approximately 1
of the adult population and 4 of children.
People who stutter (PWS) are otherwise
neurologically, psychiatrically and cognitively
unimpaired. Only a few studies have investigated
brain anatomy in PWS, revealing both structural
and functional abnormalities. One functional
abnormality characteristic of the brains of PWS
is overactivity in the cerebellum during speech
(Fig. 1). Here, we tested the hypothesis that
developmental stuttering also has a structural
correlate in the cerebellum.

• Grey Matter Volume
• The volume of the posterior lobe of the
  cerebellum was 10 smaller in the PWS group
  compared to controls (p0.04).
• This lower volume was due mainly to the RDS
  group, which was significantly different to
  controls. The difference in volume between the
  PDS group and the other two groups did not reach
  significance (Fig. 3).
• The volumes of the anterior lobes were not
  significantly different among groups.
• Across groups the right posterior lobe was bigger
  than the left but this asymmetry did not differ
  among groups.

• Fractional Anisotropy
• The FA values for the posterior lobe of the
  cerebellum were significantly higher on the right
  than the left across groups.
• In the anterior lobe, there was no significant
  difference between the PWS and the Controls. But
  there was a significant difference between the
  PDS and the RDS groups. This was due to a RgtL
  asymmetry in the PDS group, similar to the
  pattern seen in controls. The RDS group, in
  contrast, showed a LgtR asymmetry (see Fig. 4).
• Discussion
• The structural abnormalities seen here in the PWS
  compared to controls are especially striking in
  the RDS group. This is surprising given that
  their stuttering severity indices are much lower
  than the PDS group. It suggests that a greater
  degree of recovery is available to this group, in
  association with a greater degree of structural
  abnormality perhaps forcing reorganisation of
  function.
• One theory of stuttering relates to anomalous
  lateralisation of function, that is, a failure to
  show clear hemispheric dominance for language.
  The abnormal asymmetry in the WM FA values of the
  cerebellum may be related to language
  lateralisation in the PWS groups.

Figure 1
Methods

• Subjects
• 18 PWS 12 M and 6 F aged 14-27 years (avg. 18
  years). Of these, 8 had recovered from stuttering
  during their teenage years (RDS group) one of
  them was left-handed. The remaining 10 had
  persistent developmental stuttering (PDS group)
  two of them were left-handed.
• 12 age-matched controls 7 M and 5 F, aged 14-27
  years (avg. 17.8 years) all of them
  right-handed.
• Acquisitions
• Whole-head T1-weighted MRI scans, TR12, TE
  5.65, Flip angle19, 1-mm isotropic voxels, 3
  averages.
• Whole-head diffusion-weighted MRI scans, 2.5-mm
  isotropic voxels, 60 directions, 2 averages.
• Scanner
• 1.5 T - Siemens, Sonata.

• Analysis
• Cerebellar parcellation The structural images
  were linearly transformed using 12 parameters
  into standard space to aid identification of
  the primary fissure and to normalise the
  orientation of the cerebellar structures. Using
  FSLview, which displays three orthogonal planes
  simultaneously, we manually demarcated left and
  right, anterior and posterior lobes of the
  cerebellum using the midline and the primary
  fissure as landmarks (Fig. 2).
• Volume measurement Using FAST (FMRIBs
  Automatic Segmentation Tool) we calculated
  partial volume estimates for grey matter (GM) and
  white matter (WM). The partial volume maps were
  masked by the cerebellar masks to obtain GM and
  WM volumes for each region.
• Fractional Anisotropy Fractional anisotropy
  (FA) values were calculated using FDT (FMRIBs
  Diffusion Toolbox). The WM segment of the
  cerebellar lobes was used to mask the diffusion
  images and the average FA value for each of these
  masks was calculated.
• Statistical analysis Analysis of variance was
  used to compare GM and WM volumes and FA values
  between PWS (RDS and PDS groups combined and
  separately) and controls examining the factors of
  hemispheric side and lobe.

White Matter Volume There were no differences in
the WM volumes among the groups.
Conclusions
This study indicated the existence of anatomical
abnormalities in the cerebellum of PWS,
suggesting the possibility that the functional
abnormality may be related to a structural one.
It also supports the possibility that the
cerebellum plays an important role in the
pathophysiology of developmental stuttering,
perhaps in relation to lateralisation of
function.
Figure 2