Article Text

Download PDFPDF
The hummingbird identifies psp among patients with non-fluent primary progressive aphasia
  1. Matthew Silsby1,
  2. Ry Tweedie-Cullen2,
  3. Cynthia R Murray3,
  4. Glenda M Halliday3,4,
  5. John R Hodges3,4,
  6. James R Burrell1,3,4,5
  1. 1Neurology, Concord Hospital, Concord, NSW, Australia
  2. 2University of Auckland, Auckland, New Zealand
  3. 3NeuRA, Neuroscience Research Australia, Sydney, NSW, Australia
  4. 4UNSW, University of New South Wales, Sydney, NSW, Australia
  5. 5Sydney Medical School, University of Sydney, Sydney, NSW, Australia


Objectives Non-fluent primary progressive aphasias include progressive non-fluent aphasia (PNFA), a subtype of frontotemporal dementia,1 and logopenic progressive aphasia (LPA), an atypical form of Alzheimer’s disease.1–3 Progressive supranuclear palsy (PSP) overlaps with PNFA at both clinical and pathological levels; cognitive deficits are well recognised and may precede the typical motor syndrome.4 The hummingbird sign, reflected in a reduced midbrain to pons ratio, distinguishes PSP from other parkinsonian disorders. Our objective was to determine if the midbrain to pons ratio (M/P) could be used as a biomarker of PSP in patients with non-fluent primary progressive aphasia syndromes.

Methods Patients with PSP, PNFA and LPA were recruited. Patients were diagnosed clinically but pathological confirmation was available in a proportion of patients. Midbrain and pons areas were measured using Osirix Lite, a free DICOM viewer, and M/P was calculated.

Results Seventy-two participants were included (PSP 16, PNFA 18, LPA 16, and Control 22). PSP patients had motor features typical of the syndrome. M/P was significantly reduced in all disease groups, with greater reduction in PSP than in PNFA and LPA (p<0.001). An M/P ratio of less than or equal to 0.214 produced a positive predictive value of 92.9% for the diagnosis of PSP syndrome. Pathological confirmation revealed Alzheimer’s disease in 3 cases (all LPA), pathological PSP in 2 cases (1 clinical PSP and 1 PNFA) and corticobasal degeneration in 1 case (PNFA). M/P was 0.214 in the patient with clinical and pathological PSP, and 0.217 in the patient with PNFA due to PSP pathology.

Conclusions The M/P ratio was disproportionately reduced in PSP. Longitudinal studies are needed to examine the M/P ratio in patients with non-fluent aphasias in whom the PSP motor syndrome is yet to emerge. Larger studies of pathologically confirmed cases are also needed to establish M/P as a biomarker of PSP pathology.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.