Speech arrest by repetitive Transcranial Magnetic Stimulation – does it still work? Old experiences with new improvements

1.1 Clinical rationale for the study

High temporal resolution rTMS offers a non-invasive insight into speech organisation in the brain even without precise but not always available frameless stereotaxy used to target brain regions. The objective of this study was to reconsider the old methodology (Epstein et al., 1996; Pascual-Leone et al., 1991) to establish an easily available method for speech arrest based on anatomical landmarks and verify the updated procedure on a representative group of right-handed individuals.

2.1 Ethical approval

All procedures performed in studies were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The local ethical board of the Poznan University of Medical Sciences approved this study (approval No KB 523/2011) and all participants gave prior written informed consent.

2.2 Study group

Participants were included in the study if self-determined as right-handed. The group of 47 healthy adults consisted of 23 men and 24 women and included three of the paper’s authors. All participants were Polish native speakers with a median age of 23, ranging from 19 to 34 years of age.

2.3 Study design

2.4 Statistical analysis

The analysis was performed using StatSoft Statistica v10.0 and Analyse-it for Microsoft Excel v3.53 software. Methods of descriptive statistics, such as frequency (n), and arithmetic means for normally distributed variables were adopted. Otherwise, medians were used. The Shapiro–Wilk test was used to assess normality. To compare differences between the groups, the chi-square test or Fisher’s exact test were used for categorical variables, and the Mann–Whitney U test for continuous variables. Correlations between the self-assessed efficacy and comfort of stimulation and motor threshold and efficacy or comfort parameters were calculated using the Spearman rank correlation method. A p-value below 0.05 was regarded as statistically significant.

The motor cortex for orofacial, pharyngeal, laryngeal muscles and Broca’s areas are involved in the motor component of speech. The motor cortex is linked to the execution of speech production and motor speech control. Broca’s area, located in the IFG, functions as the centre of lexical, grammatical and phonological processing (Sahin, Pinker, Cash, Schomer, & Halgren, 2009). Since both the motor cortex and Broca’s area are functionally connected (Greenlee et al., 2004). IFG stimulation may induce speech disruptions.

Functional, anatomic or heuristic criteria are mostly applied to target and stimulate a chosen cortical region (Devlin & Watkins, 2007). When the functional criteria are used, participants perform a task while being stimulated (Gough, Nobre, & Devlin, 2005). Currently, neuroanatomy gives an effective method of speech centre localisation. A more precise navigated TMS involves using a traditional TMS coil fused with a frameless stereotactic MRI or CT brain scan-based online navigation system (Tarapore et al., 2013). Fiducial markers or anatomical landmarks enable localised stimulation with an error of about 5.7 mm (Ruohonen & Karhu, 2010) in navigated TMS. However, the availability of navigated TMS is still limited making it difficult to adopt in clinical neuroscience.

We decided to adopt a methodology to make the whole procedure as accessible as possible. Traditionally, rTMS coil positioning may follow the heuristic criteria using a 10–20 system (Pascual-Leone et al., 1991) or be based on the external anatomical points such as vertex, inion or tragus. We took scalp measurements (Epstein et al., 1996) as this method does not require the use of sophisticated tools.

4.2 Stimulation parameters

In the previous studies, stimulation frequencies ranged from 4 (Epstein et al., 1996, 1999, 2000), 5–18 (Wassermann et al., 1999), 16–20 (Michelucci et al., 1994), 25 (Pascual-Leone et al., 1991) to 30 Hz (Jennum et al., 1994). However, a study by Epstein and colleagues (Epstein et al., 1996) revealed that rates as low as 4 Hz are highly effective at blocking speech output, and limit the accompanying discomfort. Additionally, train length can vary from 1 second (Jennum et al., 1994) to 10 seconds (Pascual-Leone et al., 1991). In our study, we tried to minimise any possible risk and assess the efficacy of 2-second stimulation at 2 Hz. We achieved a 95.7% stimulation effectiveness which supports the application of this methodology.

There was variability across the coils used: Cadwell Rapid-Rate Magnetic Stimulator (Michelucci et al., 1994; Pascual-Leone et al., 1991), Cadwell High-Strength Magnetic Stimulator (Epstein et al., 1996; Wassermann et al., 1999), MagPro biphasic (Jennum et al., 1994), custom circular and figure-eight (butterfly) coil (Stewart et al., 2001), which was the most widely adopted variant at the time this study took place (Tarapore et al., 2013). It is recommended that the coil be oriented horizontally as then the stimulation bundle penetrates cortex areas most effectively (Epstein et al., 1996). Too high parameters of stimulator output may exceed the pain threshold and induce interpretational obstacles such as facial muscle contraction. Therefore, as recommended by Epstein (Epstein et al., 1996), we find it reasonable to use 120% of MT in the first speech arrest attempt and 150% of MT in the second attempt if necessary. Importantly, the increasing stimulator output percentage correlated significantly with discomfort assessment. Therefore, the upper limit of the stimulation values should be deemed to be 150% of MT (Mottaghy, Sparing, & Töpper, 2006).

4.3 Video recording

The video recordings enabled the precise documentation of the stimulation procedure and a non-biased analysis of its effects and positivity (Lioumis et al., 2012) by several independent reviewers. Additionally, the study volunteers could be consulted about the experiences and accompanying emotions. This enabled more thorough documentation of the experiment (Lioumis et al., 2012).

4.4 Assessment

We decided to compare the effects of speech arrest across participants by subtyping them (Thiel et al., 2005). Our methodology had the second-highest efficacy of positive speech arrest in comparison with seven previously published rTMS studies (Tarapore et al., 2013).

Interestingly, total speech arrest was observed significantly more often in participants with a speech centre active exclusively in the left hemisphere (Table 3). Participants whose speech centre was active bilaterally tended to present a subtotal stimulation effect, especially because of the effect for the right hemisphere. This is in favour of the hypothesis that speech centre activity inhibition has a greater impact on its function when the speech centre is located exclusively in the left hemisphere. Knecht et al. (Knecht et al., 2000) as discussed by Devlin and Watkins (Devlin & Watkins, 2007), presented similar findings where strongly lateralised subjects were more severely affected by unilateral TMS than those with more bilateral language organisation. Considering pre-morbid differences, the right hemisphere is more or less receptive to language before any reorganisation occurs. Essentially, the right hemisphere may play an important role in the determination of the likelihood of the right hemisphere compensation following left-sided damage.

4.5 Self-assessment

We introduced the novel procedure of speech arrest self-assessment based on a special scale, inspired by the Visual Analogue Scale used in pain assessment. The volunteers were also asked about the level of discomfort while being stimulated to consider the ratio safety/efficacy. We further suggest that the participants should be asked whether they assess the effect of stimulation as ‘Cannot recall a word’ and ‘Cannot articulate a word’ to get an insight into the subjective experience of the stimulation.

4.6 Speech arrest results

Khedr et al. (Khedr et al., 2002) investigated right-handed subjects and recorded left-sided language cerebral dominance in 87.5%, bilateral cerebral representation in 8.2%, and right-sided language cerebral dominance in 4.2%. In our study, we present the opposite results. Speech arrest was induced after stimulation over the right and left hemispheres in 57.8% and exclusively over the left hemisphere in 42.2% of volunteers. The differences in the results may result from the methodology applied. Khedr et al. (Khedr et al., 2002) considered only total speech arrest as an effect of the stimulation and used the Stanley Coren Handedness Inventory (Coren, 1993) to assess the participants’ handedness. The predominance of the bilateral positivity of speech arrest may be supported by Cogan et al. (Cogan et al., 2014) who established a bilateral sublexical speech sensory-motor system. Hickok et al. (Hickok & Poeppel, 2007) and Price et al. (Price, 2010) also demonstrated bilateral speech centre activity. The activation of the right inferior frontal cortex in language processing has been shown in both aphasic patients and healthy participants (Raboyeau et al., 2008).

There is a growing body of evidence that both hemispheres may have similar relevance to the process in the healthy brain (Hartwigsen & Siebner, 2013). The right cerebral hemisphere significantly contributes to normal language processing (Lindell, 2006). Furthermore, the right hemisphere may be either facilitating (Raboyeau et al., 2008) or deleterious (Lindell, 2006; Rosen et al., 2000) in language recovery after left IFG damage, which offers exciting possibilities in the domain of neurorehabilitation (Flöel et al., 2012).

4.7 Limitations of the methodology

Since we had limited access to advanced tools, such as navigated rTMS, when conducting this study, we were inspired by this limitation and worked towards creating a simplified, common procedure. It may be doubtful whether the activation of significantly more bilateral cortical and subcortical brain networks than previously reported results from the use of a counting task (Vanlancker-Sidtis, McIntosh, & Grafton, 2003). The coil placement was determined using the heuristic approach, which does not account for individual differences in cortical morphology. Moreover, we did not induce a speech arrest effect in two participants. We hypothesise that this may be due to increased skull thickness and decreased field permeability. However, we believe that the outcome of inducing speech arrest in over 95% of the study participants proves the reliability of our methodology.

4.8 A positive side-effect

All participants, who were doctors and medical students (including the paper’s authors), agreed that an experience of speech arrest greatly increased their understanding and empathy for people suffering from aphasia. It is an uncontested and immeasurable advantage of this cognitive research and an interesting application.

4.9 Conclusion

Repetitive Transcranial Magnetic Stimulation still constitutes an important tool in the determination of cortical speech centre activity. Involvement of the new methods enriches possible data interpretation. Our results of positive speech arrest in 95.7% of participants of a large group of 47 right-handed healthy individuals prove that the chosen stimulation parameters present a good efficacy ratio and their use in subsequent analyses seems to be justified. The traditional applications of rTMS in speech studies may be highly broadened if the methods used are further improved and simplified.

4.10 Future directions

Further studies are necessary to determine whether such results indicate that people with bilateral speech centre activity experience less extensive effects of ischemia on speech centre activity and therefore fewer instances of post-stroke aphasia. We suggest that the use of a tool such as a voice analyser might be beneficial for the objectification of speech arrest assessment in future studies.