If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Functional transcranial Doppler like functional MRI may replace language Wada test.
•
FTCD is easily applied, non-invasive and can be used for serial evaluations.
•
Language dominance shifts in women corrupt the reliability of FTCD.
•
Concerns must be raised toward single applications of non-deactivating methods.
•
The Wada seems to produce more reliable and menstrual cycle independent results.
Abstract
Purpose
Functional transcranial Doppler sonography (fTCD) is a valid and non-invasive tool for determining language dominance, e.g. in the context of presurgical evaluations. Beyond this, fTCD might be an ideal tool to study dynamics in language dominance over time. However, an essential prerequisite would be a high test–retest reliability. This was addressed in the present study.
Methods
Test–retest reliability of hemispheric hemodynamics during open speech was determined in 11 male and 11 female healthy volunteers using the Animation Description Paradigm. Expressive language dominance was assessed weekly over an interval of 4–5 weeks.
Results
Internal consistency of the four measurements was excellent (split-half reliability 0.85–0.95), but test–retest reliability of the lateralization index was poor to moderate (rtt = 0.37–0.74). Controlling for gender, test–retest reliabilities were better in men (rtt = 0.67–0.78) as compared to women (rtt = 0.04–0.70). When arranging the assessments in women around day one of menstruation – all were on contraceptives – a significant shift from left hemisphere dominance toward bilaterality (t = 2.2 p = 0.04) was evident around menstruation with significant reversal afterwards (t = −3.4 p = 0.005).
Conclusion
A high intraindividual variability of language dominance patterns is indicated in women when assessed repeatedly by fTCD. Menstrual cycle appeared to be the source of inconsistency. The finding challenges the use of non-deactivating methods for language dominance assessment in epilepsy. Support for this is demonstrated with a female patient with epilepsy in whom language dominance assessed by repeated fMRI and fTCD varied concordantly with cycle but not so the repeated intracarotidal amobarbital test.
]. This procedure, however, is invasive and today it appears mainly indicated if brain surgery may directly affect suggested eloquent cortex subserving language function. During the IAT a barbiturate is injected via a femoral catheter which terminates within the left/right arteria carotis interna. Because of its invasiveness, the IAT is neither suited for research questions beyond its strict clinical indication, nor for routinely performed follow-up assessments. Repeated measurement would for example be appreciated to answer questions regarding changes in language dominance along with language development or processes of functional recovery and plasticity in the context of chronic or reversible cerebral pathology (in epilepsy: lesions, surgery vs. seizures, interictal activity).
Among the non-invasive methods discussed as alternatives for the IAT functional MRI (fMRI) in particular meets all the criteria for utilization in clinical settings [
]. Another non-invasive alternative to the IAT is the functional transcranial Doppler sonography (fTCD), which had been introduced by Aaslid and colleagues in 1982 [
]. FTCD like fMRI assesses task-dependent changes of cerebral blood flow and is therefore an indirect measure of brain activation. During cognitive performance (e.g. word generation, picture description), the increased need for energy (glucose and oxygen) in the activated brain regions leads to increased regional blood flow that is accompanied by blood flow velocity changes within major cerebral arteries. As a result, change in neuronal activity can be indirectly assessed by experimentally measuring changes in blood flow velocity along with different mental states. FTCD is non-invasive, inexpensive, and easily applied in both children and adults. FTCD has proven to be a valid tool for the assessment of language dominance and is increasingly used in the presurgical evaluation of patients with epilepsy [
], the subject is asked to watch brief cartoon clips which have to be described in a fixed time interval immediately after presentation. The description period is followed by a silent resting phase (see Fig. 1). Different from the classical word generation paradigm, this task is easy to perform and eligible for adults, children and disabled persons as well, there is no right or wrong, the resting intervals are short, and overt speech allows for direct monitoring of the requested behavioral response. In a still ongoing validation study in patients with epilepsy the fTCD results obtained with this Animation Description Paradigm are compared to those of the IAT and/or language fMRI. Preliminary results had shown perfect correspondence between fTCD and IAT (100%, n = 10) and good concordance between fTCD and fMRI (89%; n = 27) and fMRI and IAT (90%; n = 10) [
Fig. 1Mean blood flow velocity changes within the left (red) and right (green) middle cerebral artery registered during the Animation Description paradigm (fTCD-LI: +4.0 ± 0.39).
The present study was set up to determine the internal consistency and test-retest reliability of fTCD lateralization indices in healthy subjects. Following the reliability study applying the classical silent word generation paradigm [
] the a-priori hypothesis was that the results should be stable over time. Due to its low demanding properties, we did not expect practice or habituation effects.
We did, however, control for the possibility of gender effects. In this regard there is an ongoing discussion that women may have more bilateral language representation as compared to men [
]. Apart from this, it has been demonstrated in healthy subjects that menstrual cycle causes intraindividual fluctuations in language dominance patterns as assessed by fMRI using a semantic decision task [
]. In the study by Fernandez there was greater right hemisphere activation during the luteal phase. Similarly another fMRI study showed different fMRI activation at the luteal phase and menstruation using a word generation task [
Cycle and gender-specific cerebral activation during a verb generation task using fMRI: comparison of women in different cycle phases, under oral contraception, and men.
]. Therefore, we asked all women about the time of menstruation during the study phase.
2. Methods
2.1 Participants
Twelve male and 12 female healthy volunteers were recruited to participate in this prospective longitudinal study. Past neurological psychiatric diseases led to an exclusion from the study. An inclusion criterion for women was hormonal contraceptives to eliminate fluctuations of dominance patterns in women as described by Fernandez et al. [
]. Taking this nevertheless into consideration, all female participants had to indicate the first day of their menstruation. From the original sample of 24 participants, one subject had to be excluded because of insufficient quality of the fTCD signal, probably due to a deficient temporal bone window. Another subject was not able to attend all sessions because of illness and therefore had to be excluded as well. This resulted in a total sample of 22 participants. The participantś chronological age ranged from 20 to 55 years (mean: 27.3 years; SD: 7.3). Twenty-three of the participants were right-handed and one participant was left-handed. German was the first language for all subjects. The participants took part in the study on a voluntary basis and did not receive any monetary compensation. Written informed consent was given in the first session by all participants.
2.2 Design and procedure
All subjects attended four sessions (T1–T4). In women a fifth session was added to have a session before during and after the menstruation. The assessments were carried out within four/five weeks; every week at the same day (plus minus one day). We employed the Animation Description paradigm which comprises 30 trials and takes 15 min to complete [
] (Fig. 1). Since we had found a higher concordance with language dominance results derived from IAT and fMRI, we chose the classical resting phase instead of the watching phase (as done by Bishop) as the baseline for evaluating speech activation. [
At T1 demographic data and handedness were assessed. As already mentioned female participants were also asked to indicate the first day of their menstruation along the test series.
Test sessions were as follows: The participant was placed in front of the computer. Then, the headset with the probes was attached to the participants head. If needed, glasses were put on over the fixation device. After confirmation that each probe showed the characteristic envelope curve of the middle cerebral artery (MCA), the participant had to count from one to ten to verify a stable signal during speech. Afterwards, the Animation Description paradigm was explained. The participant was asked to avoid unnecessary movements to reduce artifacts induced by muscle activation. Furthermore, the relevance of continuous speech production during the description phase was stressed.
2.3 Apparatus
The blood flow of the left and the right MCA was measured by a Doppler ultrasonography device (Doppler Box by DWL, Singen, Germany). Two 2-MHz ultrasound probes were attached to the fixation device. The stimuli were presented on a PC, using E-prime software (Psychology Software Tools).
2.4 Data analysis
The data collected from each session were analyzed using the Average 1.85 software [
]. The first step was an automated standardized post-processing of the blood flow data. Then the absolute blood flow velocities were normalized according to the baseline and filtered for outliers in the non-physiological range. After averaging the relative blood flow velocities of all accepted epochs, a mean relative blood flow velocity curve for the left and right artery as well as the lateralization index were calculated. The lateralization index (LI) describes the direction and degree of lateralization and its calculation is based on the maximum difference in cerebral blood flow velocity change between the left and right MCA during the speech versus baseline phase [
]. A positive LI indicated left hemisphere dominance, a negative LI right hemispheric dominance. If the 95% confidence interval around a given LI includes zero, a bilateral language representation is indicated.
2.5 Statistics
The statistical analyses across all subjects was performed across 4 sessions using SPSS 19. Analyses separate for men and women took 4 versus 5 tests into consideration. Split-half reliability was calculated to verify consistency across the up to 30 trials that underly the lateralization index. Test-retest reliabilities (Pearson product-moment correlations) were calculated for all combinations of assessments (T1–T5).
The within-subject stability was evaluated with repeated measures ANOVA. Furthermore, repeated measures ANOVA with gender as between-subject factor and retests as within-subject factor was calculated. Afterwards, men and women were analyzed separately. Finally, in women, menstruation was taken into account by using paired t-tests.
3. Results
Fig. 1 depicts a typical change in regional cerebral blood flow velocity that was initiated by the Animation Description paradigm.
An ANOVA did not reveal significant differences across all LIs for the factor gender (F = 0.18 p = 0.197).
Split-half reliability was calculated across all subjects for each session (T1–T5) using the Spearman-Brown formula. The results indicated reliability coefficients between 0.85 and 0.95 for the individual sessions (Table 1).
Table 1Split-half reliability coefficients for each session.
The test–retest reliability results for all subjects are shown in Table 2(a) . Although the results reached significance for each combination, the coefficients were moderate (rtt = 0.37–0.74) rather than high. Table 2(b and c) shows the test-retest reliabilities split up for gender. The results in men revealed high reliabilities (rtt = 0.67–0.78) between all sessions, this being confirmed by repeated measures ANOVA which did not reveal a significant impact of the factor time. (F = 0.05, p = 0.98). The men's mean LIs ranged from 1.4 to 1.78. (Fig. 2a )
Table 2Pearson's reliability coefficients.
T1
T2
T3
T4
T5
(a) All subjects
T1
–
0.45 (0.02)
0.41 (0.03)
0.37 (0.04)
–
T2
–
0.74 (<0.01)
0.41 (0.02)
–
T3
–
0.64 (<0.01)
–
(b) Men
T1
–
0.71 (0.01)
0.67 (0.02)
0.76 (0.01)
–
T2
–
0.74 (0.01)
0.72 (0.01)
–
T3
–
0.78 (0.01)
–
(c) Women
T1
0.27 (0.201)
0.14 (0.333)
0.04 (0.454)
−0.08 (0.797)
T2
0.70 (0.006)
0.24 (0.231)
0.18 (0.589)
T3
0.50 (0.068)
0.27 (0.421)
T4
−0.22 (0.615)
Correlation coefficients and p-values (one-tailed) in brackets.
Fig. 2Lateralization indices (a) across all sessions (men and women) (b) across sessions broken down for men and women (c) the three sessions before, next to, and after menstruation in women.
In women, except for one combination (T2–T3), none of the test-retest coefficients reached statistical significance indicating insufficient reliability over time (Table 1). Repeated measures ANOVA with 5 tests revealed no statistical significance (F = 1.47 p = 0.22). Women showed mean LI's in the range from 1.9 to 3.2. (Fig. 2b).
Following hints from the literature that changes of language dominance in women may follow hormonal variation across the menstrual cycle, the women's session order was arranged around the time of the reported menstruation. The session closest to time of the first day of menstruation was defined as well as one session the week before and one week afterwards (Fig. 2c). Whether the time point of assessment within the menstrual cycle has an influence on language lateralization was tested using paired t-tests. The results indicated that the session around menstruation differed significantly from the session before (t = 2.19; p = 0.047) and after (t = −3.383; p = 0.005), i.e. the lateralization index around menstruation showed a reversible trend toward bilaterality). The remaining two sessions did not significantly differ from each other (t = −0.811; p = 0.432). Additional reliability calculations showed that the sessions one week before and one week after the time of testing during menstruation showed a significant correlation (r = 0.45 p = 0.05) whereas the tests before and after menstruation did not correlate with the test during menstruation (pre/during r = 0.20, p = 0.244, during/post r = 0.19 p = 0.254).
4. Discussion
In times when the IAT is discussed to be no longer justified for routine assessment of patients undergoing epilepsy surgery, alternative non-invasive tools for language dominance assessment gain increasing interest. At our department fTCD and fMRI are currently used in patients with epilepsy and repeated fTCD seems particularly attractive for the monitoring of language development and processes of functional plasticity along with brain damage or surgery. Preliminary data from an ongoing cross-sectional validation study had shown good concordance of fTCD using the Animation Description Paradigm with language fMRI, and perfect concordance between fTCD and IAT [
]. The original aim of this study was the assessment of the reliability of expressive language dominance results derived from fTCD, when applying the Animation Description paradigm which allows for open speech. Stability of fTCD results is a prerequisite for longitudinal follow-up assessments.
Until now test-retest reliability of fTCD in children or adults has mainly been demonstrated by the stability across two assessments [
In the present study, language dominance was assessed weekly over 4–5 weeks. Test-retest reliability of the lateralization indices across all subjects was poor to moderate with reliability coefficients ranging from 0.37 to 0.74. Lack of reliability was mainly caused by the great variability seen in women. In men stable results with satisfactory to high reliability were obtained (rtt = 0.67–0.78). This observation raised the question whether fundamental differences between men and women led to this difference and whether such a measure can really be used as IAT replacement in patients with epilepsy.
As indicated in the introduction, we took into consideration that women might show different language lateralization than men. Averaging the lateralization index across all sessions both groups were left dominant. As discussed in the literature the sex difference may be less an absolute one than one which depends on the time of testing [
]. Language dominance as a function of menstrual cycle was not the core issue of this study, but dominance changes along the menstrual cycle obviously need to be discussed in order to understand the results obtained here. We had simply asked for the time of menstruation and did neither assess hormones, nor did we intend to contrast the luteal phase with menstruation. Post hoc arrangement of the women's assessments in relation to the assessment closest to day one of the menstruation resulted in a distinctive pattern which indicated less marked left hemispheric dominance during this time, i.e a shift toward bilaterality. Before and after the assessment at menstruation the lateralization indices clearly indicated left dominance. Previous studies [
] indicated more bilateral language and worse mental rotation in the midluteal as compared to the menstruation phase. The gonadal hormones estradiol and progesterone have been claimed to be responsible for interhemispheric coupling and decoupling [
]. During low levels of estradiol and progesterone the left hemisphere seems to inhibit the right hemisphere. In contrast during high levels of estradiol and progesterone the inhibition is diminished leading to a more symmetric cerebral function. Different from these two studies, the women in our study were all on contraceptives. According to theoretical considerations contraceptives eliminate hormonal peak levels due to the artificial harmonization [
Cycle and gender-specific cerebral activation during a verb generation task using fMRI: comparison of women in different cycle phases, under oral contraception, and men.
]. This plus the vague timing of the assessments in regard to the menstrual cycle may account for the different observations.
The design of this study clearly does not allow any in-depth analysis of the observation that menstrual cycle corrupted the reliability of language fTCD in women. In conjunction with the studies just mentioned, four major questions arise:
1.
First one would like to know the driving factor behind the variation seen in women. Is it hormonal change or a biorhythm and why is the language activation shift not seen in all women?
2.
Second, and this is maybe more important, one would like to know the behavioral consequences of such shifts. In clinical settings (i.e. lesion driven) atypical language dominance has cognitive consequences and these are in part different for women and men. In left lateralized epilepsies, atypical language dominance and preservation of language often goes along with so called crowding or suppression of originally right hemispheric functions. This can cause a pattern of material-specific cognitive impairments which is misleading in regard to the typical left vs. right and verbal vs. nonverbal dichotomy [
Sex differences in material-specific cognitive functions related to language dominance: an intracarotid amobarbital study in left temporal lobe epilepsy.
]. In this regard further research should address menstrual-cycle dependent changes in language dominance parallel to other cognitive domains, whilst taking into account hormone levels of naturally cycling women as well as contraceptive using women.
3.
The third question regards the consequences for the interpretation of findings from previous neuroimaging studies using single non-deactivating assessments of language dominance. The present data indicate that dependent on which phases of the menstrual cycle are included, findings based on a single assessment in women as a group may sum up to more bilateral language dominance patterns. In single cases even false right dominance might be indicated. Whether a false lateralization into the left hemisphere is possible in genuine right dominant women would be interesting to see.
4.
In clinical settings, i.e. in the rare conditions where it is still indicated, the results of the IAT still guide surgical interventions and the indication of additional, in part invasive diagnostic procedures (e.g. electrocortical stimulation via implanted electrodes or intraoperatively) [
]. Different from fTCD, fMRI and most other non-deactivating procedures which derive language dominance patterns from the relative contribution of both interacting hemispheres, IAT and transcranial magnetic stimulation (TMS) conclude language dominance from deactivation of the suggested language relevant hemisphere or brain region, respectively. In this regard it would also be essential to know whether results from deactivating procedures do fluctuate with menstrual cycle.
In this respect a single case report may provide first evidence of the possibility of a principal and clinically highly relevant difference between functional imaging and deactivating methods. A 28-year-old female patient suffered from chronic uncontrolled epilepsy with complex partial and infrequent secondary generalized seizures originating from the right frontal lobe. MRI post-processing (MAP 07) revealed a suspicious right frontal/insular lesion most likely a cortical dysplasia. Because of the localization of the lesion in the vincinity of the homologue to Boca's area and because of aphasic symptoms during/after the seizure, the patient underwent fTCD for screening of language dominance. The result indicated left hemispheric dominance for expressive language dominance. Since this is part of an ongoing validation study, she also underwent language fMRI three days later which also indicated left hemispheric dominance. It was decided to go for surgery but not without an IAT for final evidence. The patient came back three months later and underwent a unilateral right IAT. Language performance when testing the isolated left hemisphere, however, revealed that she was able to show only 38–40% of the baseline performance. This was interpreted as evidence that the right hemisphere might be involved in language. Since this was discordant with the imaging data (fMRI and fTCD) these tests were repeated in the following two days. Most astonishingly both tests now indicated bilateral language dominance consistent with the IAT. Knowing from the ongoing study on test–retest reliability, and having checked for the test dates in regard to the patient's menstrual cycle (1st assessment during menstruation phase, 2nd assessment during luteal phase; no intake of hormonal contraceptives; cf. [
]), the question was raised whether we can trust the IAT. Moreover, what would the IAT have shown at the time of the first functional assessments that indicated typical left hemispheric dominance? Therefore, the assessments were repeated at about the same time of the menstrual cycle when the first assessments had been conducted, i.e. during menstruation phase. Both functional imaging data consistently indicated clear left dominance. However, the right-sided IAT for testing the isolated left hemisphere, which was done one day later, once again indicated an atypical pattern with a participation of the right hemisphere in language function. At this time no expressive and only part of receptive functions were possible with the isolated left hemisphere. The intraindividual language dominance results are illustrated in Fig. 3.
Fig. 3Single case study of a patient with right frontal dysplasia who, along the menstrual cycle, demonstrated variation of language dominance when repeating non-deactivating procedures (fTCD and fMRI: 1st test time left dominant, 2nd test time bilateral, 3rd test time left dominant) and right hemisphere involvement when repeating deactivating (right sided) IAT (2nd test time: expressive and receptive, 3rd test time primarily expressive).
In concluding the present study deactivation methods seem to assess more basic and stable patterns of language dominance, irrespective of menstrual cycle. Results from activation tasks in contrast appear susceptible to dominance changes along with menstrual cycle. For clinical use and also for research questions, single measurements in women using tools that solely assess brain activation patterns of the two interacting hemispheres are thus not sufficient to provide reliable results. The study hopefully initiates follow up evaluations, since it raised more questions than having given the expected answers.
Conflict of interest statement
None of the authors has any conflict of interest in regard to the topic and the contents of this article.
References
Wada J.
Rasmussen T.
Intracarotid injection of sodium amytal for the lateralization of cerebral speech dominance. 1960.
Cycle and gender-specific cerebral activation during a verb generation task using fMRI: comparison of women in different cycle phases, under oral contraception, and men.
Sex differences in material-specific cognitive functions related to language dominance: an intracarotid amobarbital study in left temporal lobe epilepsy.
☆One of the authors of this paper is a member of the current editorial team of Seizure. The supervision of the independent peer review process was undertaken and the decision about the publication of this manuscript were made by other members of the editor.
00038-2&id=gr1.jpg){kind=link}
00038-2&id=gr2.jpg){kind=link}
00038-2&id=gr3.jpg){kind=link}