Reproducibility of measures of neurophysiological activity in Wernicke's area: A magnetic source imaging study
Introduction
High-resolution, non-invasive studies of the neuroanatomical basis of language function have been made possible in the past decade with advances in imaging techniques. Reported findings have generally been consistent with classic language models (Wernicke, 1874) that were based largely on the study of linguistic deficits in patients with brain lesions (Damasio, 1998, Price, 2000) and the results of invasive brain mapping procedures (Lesser et al., 1994, Ojemann et al., 1989). An anterior left hemisphere speech region, classically defined as Broca's area and located in the posterior part of the inferior frontal gyrus, has been shown to be active during expressive language tasks. Activation in Wernicke's area, located in the posterior portion of the superior temporal gyrus in the left hemisphere, is found during receptive language tasks. Agreement between the outline of functional images and prior expectations regarding the outline of the brain mechanism responsible for language functions instigated several clinical trials of functional brain imaging methods as tools for the identification of eloquent cortex prior to surgical resection of tumors, epileptogenic foci, and vascular malformations (e.g. Binder et al., 1997, Stapleton et al., 1997, Bookheimer et al., 1997, Simos et al., 1999a, Castillo et al., 2001, Kober et al., 2001, Rutten et al., 2002).
Given the practical and technical limitations associated with invasive presurgical brain mapping procedures (e.g. Önal et al., 2003), particularly when evaluating pediatric patients, clinicians have welcomed the availability of non-invasive brain imaging alternatives (e.g. Forss et al., 2000, Pataraia et al., 2002). Whole-head magnetoencephalography (MEG) is the newest of functional brain imaging techniques (e.g. Hämäläinen et al., 1993, Kaufman and Ju, 2003), and only recently suitable activation protocols have been developed for identifying the location and extent of language specific cortex. Those centers employing MEG for the presurgical mapping of language functions have shown very favorable results in terms of both its reliability and validity. As would be expected, language-related activity sources have been extracted from the late components of event-related magnetic field records (ERFs), which occur later than 200 ms after the presentation of the stimulus (Halgren et al., 2002, Helenius et al., 2002, Kober et al., 2001, Simos et al., 1998, Simos et al., 1999a, Simos et al., 1999b).
A further prerequisite for the establishment of any brain imaging technique as clinically useful and equivalent to invasive techniques is that the random, intra-participant variability in source localization for selected activation protocols be substantially smaller than the spatial resolution limit of the gold-standard method which is in the order of 1 cm. While reproducibility studies have been successfully completed for early sensory (Baumann et al., 1990, Gallen et al., 1994, Pantev et al., 1991) and motor-related activity (Castillo et al., 2004), there is a scarcity of studies investigating the test–retest reliability of late components associated with language-specific neurophysiological operations (Breier et al., 2000).
In the present study, we examine the reproducibility of estimates of magnetic activity associated with the engagement of Wernicke's area in a simple word recognition task in healthy volunteers. The activation protocol used in this study is identical to the one used in previous clinical validation studies, and it is known to produce reliable magnetic activity sources in posterior temporal and temporoparietal regions in the left hemisphere, where Wernicke's area is located in the vast majority of neurologically intact right-handers. These sources represent cortical electrical activity that takes place during the late phases of stimulus processing (i.e. >200 ms after stimulus onset) and lasts for several hundred milliseconds. This activation protocol was, therefore, deemed suitable for studying the stability of estimates of language-specific, regional brain activity. Reproducibility measures were based on intrasession (split-half) variability, affected by short-term changes in state factors influencing regional neurophysiological activation (motivation, attention, fatigue, habituation, etc.), as well as by random changes in extraneous and biological magnetic noise.
Section snippets
Subjects
Participants were 14 right-handed healthy volunteers (seven men) between 24 and 52 years of age (mean: 33.4 years) without history of neurological or psychiatric disease, or hearing impairment. Hand preference was evaluated using the Edinburgh Handedness Inventory (Oldfield, 1971) with all subjects scoring +0.6 or greater. Prior to the study all participants gave their written informed consent, after the purpose of the study and the nature of the procedures involved had been explained to them.
Stimuli and task
Preliminary analyses
In every dataset magnetic flux during the 150 ms pre-stimulus interval, averaged across all magnetic sensors used to calculate the Wernicke's area ECDs (an estimate of background environmental and biological noise), was below 10 fT. Field strength for ECDs comprising each cluster ranged between 31 and 205 fT across participants (i.e. over three times higher than background noise), a level that is generally considered adequate for reliable computation of ECD source parameters. The vast majority of
Discussion
Previous studies investigating the test–retest reliability of MSI focused exclusively on records of early activity originating from modality-specific sensory (auditory or somatosensory) cortices (Gallen et al., 1994, Pantev et al., 1991). The present study is the first to our knowledge of the reproducibility of MEG-derived estimates of late activity in the context of a language task. Moreover, the activation task and data acquisition and analyses procedures used in the present study were
Acknowledgements
This study was supported in part by NINDS grant NS37941 and NICHD/OERI grant HD30885to Andrew C. Papanicolaou, and by the Vivian L. Smith Center for Neurologic Research, Houston, TX.
References (45)
- et al.
Integrating sensory and motor mapping in a comprehensive MEG protocol. Clinical validity and replicability
Neurolmage
(2004) - et al.
Intra-subject replication of brain magnetic activity during the processing of speech sounds
Cogn Brain Res
(2004) - et al.
Clinical use of magnetoencephalography
Suppl Clin Neurophysiol
(2000) - et al.
Intrasubject reliability and validity of somatosensory source localization using a large array biomagnetometer
Electroencephalogr Clin Neurophysiol
(1994) - et al.
N400-like magnetoencephalography responses modulated by semantic context, word frequency, and lexical class in sentences
Neuroimage
(2002) - et al.
A study of dipole localization accuracy for MEG and EEG using a human skull phantom
Electroencephalogr Clin Neurophysiol
(1998) - et al.
Spanish language mapping using MEG: a validation study
Neuroimage
(2002) The assessment and analysis of handedness: the Edinburgh inventory
Neuropsychologia
(1971)- et al.
The 3D topography of MEG source localization accuracy: effects of conductor model and noise
Clin Neurophysiol
(2003) - et al.
Replicability of MEG and EEG measures of the auditory N1/N1m-response
Electroencephalogr Clin Neurophysiol
(1998)