Intrasubject Reproducibility of Functional MR Imaging Activation in Language Tasks
G.S. Harringtona,
M.H. Buonocorea and
S. Tomaszewski Fariasb
a Department of Radiology, University of California at Davis, Davis, Calif
b Department of Neurology, University of California at Davis, Davis, Calif

View larger version (61K):
[in a new window]
|
Fig 1. Reproducibility for fMRI activation: LI and volume.
Top row, The LI and test-retest correlation of LI for all tasks and task combinations within both regions of interest. Middle row, The average volume of activation and % subjects with activation during both imaging sessions within the language dominant hemisphere for both regions of interest. Bottom row, Average concurrence ratios (Rvolume and Roverlap) within the language dominant hemisphere for both regions of interest. SD = semantic decision; CN = confrontation naming; VG = verb generation; AS = audio sentence comprehension; VS = visual sentence comprehension; SL = story listening; CTa = combination of SD, CN, and VG; CTb = combination of AS,VS, and SL; Ctc = combination of all tasks.
| |

View larger version (114K):
[in a new window]
|
Fig 2. Verb-generation overlap.
Activation for the verb-generation task in 2 example subjects. Images are in radiologic view. Yellow = session 1 only; red = session 2 only; blue = overlap between sessions. Activation information provided in the Table.
| |

View larger version (16K):
[in a new window]
|
Fig 3. Temporal SNR versus max head motion. Regression analyses: tSNRmed: r = 0.42; P = 2.2 x 106; tSNRlow: r = 0.14; P = .13. White = tSNRlow; black = tSNRmed.
| |