Neuropediatrics 2019; 50(S 02): S1-S55
DOI: 10.1055/s-0039-1698243
Poster Presentations
Poster Area GNP Varia 2/Stroke
Georg Thieme Verlag KG Stuttgart · New York

Speech fMRI in Children, Adolescents and Young Adults: Identification and Evaluation of an Optimal Study Design and Analysis Algorithm

Hannah Krafft
1   Schön Klinik Vogtareuth, Center for Pediatric Neurology, Neurorehabilitation and Epileptology, Vogtareuth, Germany
2   University of Tübingen, Children’s Hospital, Department of Pediatric Neurology and Developmental Medicine, Tübingen, Germany
,
Tom Pieper
3   Schön Klinik Vogtareuth, Klinik für Neuropädiatrie und Neurologische Rehabilitation, Epilepsiezentrum für Kinder und Jugendliche, Vogtareuth, Germany
,
Manfred Kudernatsch
4   Schön Klinik Vogtareuth, Klinik für Neurochirurgie und Epilepsiechirurgie, Vogtareuth, German
,
Wiebke Hofer
3   Schön Klinik Vogtareuth, Klinik für Neuropädiatrie und Neurologische Rehabilitation, Epilepsiezentrum für Kinder und Jugendliche, Vogtareuth, Germany
,
Martin Staudt
4   Schön Klinik Vogtareuth, Klinik für Neurochirurgie und Epilepsiechirurgie, Vogtareuth, German
5   University of Tübingen, Children’s Hospital, Department of Pediatric Neurology and Developmental Medicine,Tübingen, Germany
› Author Affiliations
Further Information

Publication History

Publication Date:
11 September 2019 (online)

Also available at
 

Aim of the Study: The aim of this study was the development of a clinically practicable test and analysis strategy for speech fMRI in pediatric patients.

Material and Methods: Between 01.01.2008 and 01.04.2016, at our clinic, 172 speech-fMRIs were performed. Among these, 152 speech-fMRIs (in 131 patients, aged 5–28 years) presented in the frame of at least one speech task an activation of at least one region of interest (ROI) as previously established in a validation study (Meinhold et al, submitted) (1–3). Based on this, a hemisphere was defined as relevant for speech function if at least 2 valid ROIs were activated in at least 2 fMRI sessions (either using the same or different speech tasks).

Results: Based on these criteria, we defined language-relevant hemispheres in 104 fMRI in 90 patients. In total, language function was considered left-dominant in 66 patients, right-dominant in 7 patients and bilateral in 17 patients. The speech productive tasks, i.e., word chain task (WCT), vowel identification task (VI) or synonym task (SYN), activated frontal regions (middle frontal gyrus or inferior frontal gyrus or frontal operculum) in ≥ 82% of all sessions. WCT also produced strong activation (84%) of SMA (supplementary motoric area). 97% of all BST sessions activated the primary auditory cortex (A1) (Heschl’s gyri).

Discussion: Since all speech tests revealed areas with very frequent activation (82% frontal for productive speech tasks, 84% SMA for WCT, 97% A1 for beep stories), these regions can be used as reliable markers for successful tests. Combining these established task ROIs and using reproducibility by area or task, respectively, a validated estimation of hemispheric language dominance can be achieved and an optimized fMRI analysis can be performed.

Conclusion: Using this introduced visual fMRI analysis method, pediatric patients with clear left dominant fMRI results may no longer undergo invasive language assessment using WADA. However, further evaluations are necessary for study design optimization and validation of analysis algorithms.

References

1. Meinhold, T., et al. (see GNP-FV34 in this supplement). “Pediatric presurgical language fMRI: a validation study.”

2. Zsoter A, Staudt M, Wilke M. Identification of successful clinical fMRI sessions in children: an objective approach. Neuropediatrics 2012;43(5):249–257