| Author |
 Topic  |
|
|
Seniha
BIAC Alum
6 Posts |
 Posted - Jul 31 2003 : 11:59:40 AM
|
Dear all,
Keith and I were working on creating group average tmaps using the script that we acquired from other BIAC users. The script basically adds the tmaps of individual subjects and divides it by the Square root of n, which results in large values of t. When we display the group average t-map, almost the entire brain lights up.
We were wondering if this formula is statistically valid given that the average t-value is higher than any of the individual subject? Does anybody have a better understanding of where this formula came from?
Thanks for the help,
Seniha
|
|
|
kap
BIAC Faculty
4 Posts |
Posted - Sep 30 2003 : 7:01:28 PM
|
Hi - The formula came from this paper (in a direct kind of way) but has been around since R.A. Fisher:
Lazar NA, Luna B, Sweeney JA, Eddy WE: Combining Brains: A survey of methods for statistical pooling of information. Neuroimage 16, 538-550, 2002.
and has been used in the following BIAC papers (so I sure hope it's valid... if not, please don't tell anybody!):
1. Pelphrey, K. A., Singerman, J. D., Allison, T., & McCarthy, G. (2003). Brain activation evoked by the perception of gaze shifts: The influence of context. Neuropsychologia, 41, 156-170.
2. Pelphrey, K. A., Mitchell, T. V., McKeown, M., Goldstein, J., Allison, T., & McCarthy, G. (2003). Brain activity evoked by perception of human walking: Controlling for meaningful coherent motion. Journal of Neuroscience, 23, 6819-6825.
3. Wright, T. M., Pelphrey, K. A., Allison, T., & McCarthy, G. (2003). Polysensory interactions evoked by audiovisual speech perception along the superior temporal sulcus region. Cerebral Cortex, 13, 1034-1043.
4. Pelphrey, K. A., Viola, R. J., & McCarthy, G. (2003, in press). When strangers pass: Processing of mutual and averted gaze in the superior temporal sulcus. Psychological Science.
The idea is that you are combining statistics using a simple fixed effects analysis – so you have to account for this when you divide. We can debate the relative merits and demerits of fixed versus random effects. Really it depends on what inferences you want to make - the opinion of some British neuroimagers aside. Think of it like a meta-analysis in which each fMRI subject is a study. In fact, the formula originally came from the meta-analysis literature. See for example, the Handbook of Research Synthesis - Cooper, Hedges - 1994.
By the way, I am not endorsing this method as a standard. Each person should consider the question they want to address, read Luna's paper about the options and papers by others and make an informed decision.
- Kevin
|
 |
|
| |
Topic |
|
|
|
creating group average t-maps
