PURPOSE: To investigate intersite variability of clinical functional magnetic resonance (MR) imaging, including influence of task standardization on variability and use of various parameters to inform the clinician whether the reliability of a given functional localization is high or low. MATERIALS AND METHODS: Local ethics committees approved the study; all participants gave written informed consent. Eight women and seven men (mean age, 40 years) were prospectively investigated at three experienced functional MR sites with 1.5- (two sites) or 3-T (one site) MR. Nonstandardized motor and highly standardized somatosensory versions of a frequently requested clinical task (localization of the primary sensorimotor cortex) were used. Perirolandic functional MR variability was assessed (peak activation variability, center of mass [COM] variability, intraclass correlation values, overlap ratio [OR], activation size ratio). Data quality measures for functional MR images included percentage signal change (PSC), contrast-to-noise ratio (CNR), and head motion parameters. Data were analyzed with analysis of variance and a correlation analysis. RESULTS: Localization of perirolandic functional MR activity differed by 8 mm (peak activity) and 6 mm (COM activity) among sites. Peak activation varied up to 16.5 mm (COM range, 0.4-16.5 mm) and 45.5 mm (peak activity range, 1.8-45.5 mm). Signal strength (PSC, CNR) was significantly lower for the somatosensory task (mean PSC, 1.0% ± 0.5 [standard deviation]; mean CNR, 1.2 ± 0.4) than for the motor task (mean PSC, 2.4% ± 0.8; mean CNR, 2.9 ± 0.9) (P < .001, both). Intersite variability was larger with low signal strength (negative correlations between signal strength and peak activation variability) even if the task was highly standardized (mean OR, 22.0% ± 18.9 [somatosensory task] and 50.1% ± 18.8 [motor task]). CONCLUSION: Clinical practice and clinical functional MR biomarker studies should consider that the center of task-specific brain activation may vary up to 16.5 mm, with the investigating site, and should maximize functional MR signal strength and evaluate reliability of local results with PSC and CNR.
PURPOSE: To investigate intersite variability of clinical functional magnetic resonance (MR) imaging, including influence of task standardization on variability and use of various parameters to inform the clinician whether the reliability of a given functional localization is high or low. MATERIALS AND METHODS: Local ethics committees approved the study; all participants gave written informed consent. Eight women and seven men (mean age, 40 years) were prospectively investigated at three experienced functional MR sites with 1.5- (two sites) or 3-T (one site) MR. Nonstandardized motor and highly standardized somatosensory versions of a frequently requested clinical task (localization of the primary sensorimotor cortex) were used. Perirolandic functional MR variability was assessed (peak activation variability, center of mass [COM] variability, intraclass correlation values, overlap ratio [OR], activation size ratio). Data quality measures for functional MR images included percentage signal change (PSC), contrast-to-noise ratio (CNR), and head motion parameters. Data were analyzed with analysis of variance and a correlation analysis. RESULTS: Localization of perirolandic functional MR activity differed by 8 mm (peak activity) and 6 mm (COM activity) among sites. Peak activation varied up to 16.5 mm (COM range, 0.4-16.5 mm) and 45.5 mm (peak activity range, 1.8-45.5 mm). Signal strength (PSC, CNR) was significantly lower for the somatosensory task (mean PSC, 1.0% ± 0.5 [standard deviation]; mean CNR, 1.2 ± 0.4) than for the motor task (mean PSC, 2.4% ± 0.8; mean CNR, 2.9 ± 0.9) (P < .001, both). Intersite variability was larger with low signal strength (negative correlations between signal strength and peak activation variability) even if the task was highly standardized (mean OR, 22.0% ± 18.9 [somatosensory task] and 50.1% ± 18.8 [motor task]). CONCLUSION: Clinical practice and clinical functional MR biomarker studies should consider that the center of task-specific brain activation may vary up to 16.5 mm, with the investigating site, and should maximize functional MR signal strength and evaluate reliability of local results with PSC and CNR.
Authors: Dylan G Gee; Sarah C McEwen; Jennifer K Forsyth; Kristen M Haut; Carrie E Bearden; Jean Addington; Bradley Goodyear; Kristin S Cadenhead; Heline Mirzakhanian; Barbara A Cornblatt; Doreen Olvet; Daniel H Mathalon; Thomas H McGlashan; Diana O Perkins; Aysenil Belger; Larry J Seidman; Heidi Thermenos; Ming T Tsuang; Theo G M van Erp; Elaine F Walker; Stephan Hamann; Scott W Woods; Todd Constable; Tyrone D Cannon Journal: Hum Brain Mapp Date: 2015-03-28 Impact factor: 5.038
Authors: Jakob Rath; Moritz Wurnig; Florian Fischmeister; Nicolaus Klinger; Ilse Höllinger; Alexander Geißler; Markus Aichhorn; Thomas Foki; Martin Kronbichler; Janpeter Nickel; Christian Siedentopf; Wolfgang Staffen; Michael Verius; Stefan Golaszewski; Florian Koppelstaetter; Eduard Auff; Stephan Felber; Rüdiger J Seitz; Roland Beisteiner Journal: Hum Brain Mapp Date: 2016-03-09 Impact factor: 5.038
Authors: Eric C Leuthardt; Maurizio Corbetta; Carl D Hacker; Timothy O Laumann; Nicholas P Szrama; Antonello Baldassarre; Abraham Z Snyder Journal: Neuroimage Date: 2013-06-02 Impact factor: 6.556
Authors: A Geißler; E Matt; F Fischmeister; M Wurnig; B Dymerska; E Knosp; M Feucht; S Trattnig; E Auff; W T Fitch; S Robinson; R Beisteiner Journal: Neuroimage Date: 2014-09-22 Impact factor: 6.556
Authors: Melanie A Morrison; Nathan W Churchill; Michael D Cusimano; Tom A Schweizer; Sunit Das; Simon J Graham Journal: PLoS One Date: 2016-02-19 Impact factor: 3.240