Activity Number:
|
367
|
Type:
|
Invited
|
Date/Time:
|
Thursday, August 15, 2002 : 8:30 AM to 10:30 AM
|
Sponsor:
|
Biopharmaceutical Section*
|
Abstract - #301637 |
Title:
|
What Can Modern Statistics Offer Imaging Neuroscience?
|
Author(s):
|
Carrie Wager*+ and Nicholas Lange+
|
Affiliation(s):
|
Harvard School of Public Health and Harvard School of Public Health
|
Address:
|
655 Huntington Avenue, Boston, Massachusetts, 02115, USA 677 Huntington Avenue, Boston, Massachusetts, 02115, USA
|
Keywords:
|
magnetic resonance imaging ; microscopy ; gene expression ; kriging ; generalized linear mixed model ; mathematical morphology
|
Abstract:
|
This talk will begin with a brief presentation of two current problems in human and non-human animal brain imaging: human pharmacological MRI and spatial intensity estimation in rat brain expression of a single gene. Next, several statistical analyses of each study will be discussed. Each analysis involves modern statistical methodology, yet each also contains standard classical elements. The main goal will then be to place these statistical analyses squarely within their neuroscientific contexts and ask: Do the modern statistical approaches and results help to answer the primary neuroscientific questions driving these studies? In some aspects, simple traditional statistical approaches may suffice; in others, modern statistical methodology may be required. We will then identify important commonalities and differences between neuroscientific and statistical viewpoints when assessing brain imaging studies and their findings in light of their importance in the broader scientific context within which they reside.
|