Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a method

YEBEH-05824; No of Pages 8 Epilepsy & Behavior xxx (2018) xxx–xxx

Contents lists available at ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a method Gunilla Rejnö-Habte Selassie a,⁎, Göran Pegenius b, Gerd Viggedal c, Tove Hallböök c,1, Magnus Thordstein b,1,2 a b c

Unit of Pediatric Speech and Language Pathology, Queen Silvia Children's Hospital, Gothenburg, Sweden Unit of Clinical Neurophysiology, Sahlgrenska University Hospital, Gothenburg, Sweden Department of Pediatrics, Queen Silvia Children's Hospital and Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

a r t i c l e

i n f o

Article history: Received 26 April 2018 Revised 20 May 2018 Accepted 20 May 2018 Available online xxxx Keywords: Transcranial magnetic stimulation Cortical language mapping Children Epilepsy

a b s t r a c t We adjusted an object-naming task with repetitive navigated transcranial magnetic stimulation (rnTMS) originally developed for preoperative cortical language mapping in adults in order for it to be used in children. Two series of pictures were chosen for children above and below 10 years of age, respectively. Firstly, the series of pictures and the preferred speed of presentation were assessed for their applicability in children of different ages and abilities. Secondly, these series were used with rnTMS preoperatively in five children with epilepsy. Naming errors induced by the stimulation comprised no response, delayed response, semantic error, phonological error, and self-correction. Language laterality was compared with the results of a dichotic listening test and with neuropsychological tests with respect to general laterality, and general language abilities were considered with respect to the results of stimulation. One participant had below normal general language abilities, two had below-normal rapid naming, and three had slow and indistinct articulation. Laterality was only clear in two of the participants. All children required breaks of various durations during the process, and individual adjustments of the interpicture interval and other stimulation parameters were also made. We conclude that, after adjustment, rnTMS combined with an object-naming task can be useful for preoperative language mapping in children. © 2018 Elsevier Inc. All rights reserved.

1. Introduction Repetitive navigated transcranial magnetic stimulation (rnTMS) has become a widely used method for preoperative cortical mapping of various abilities including motor function and expressive language. Several reports have been published on the use of this technique in adult patients due to neurosurgical operations for different conditions, mainly tumors and epilepsy [1–6]. A few studies using the same method for cortical mapping of motor ability in children are also available [7,8]. Object naming is a sensitive and clinically useful tool for cortical mapping of expressive language and is used for identifying language laterality or language localization [9,10]. Originally, intraoperative electrical cortical stimulation was used for language mapping. When conducting a transcranial magnetic stimulation for language mapping, a picture slide show is displayed on a monitor. Participants first name the objects without stimulation to establish a baseline of competence,

⁎ Corresponding author at: Unit of Pediatric Speech and Language Pathology, Queen Silvia Children's Hospital, Folkungagatan 16, 411 02 Gothenburg, Sweden. E-mail address: [email protected] (G. Rejnö-Habte Selassie). 1 Co-last authors. 2 Present address: Unit of Clinical Neurophysiology, Linköping University Hospital and Department of Clinical and Experimental Medicine, Linköping, Sweden.

and pictures which result in naming errors are discarded. The remaining pictures are shown in a random order during stimulation, and naming errors induced by the stimulation are identified. Corina et al. proposed an 11-point coding scheme for classifying naming errors by type [11]: semantic error, phonological error, grammatical error, circumlocution, neologism, perseveration, response delay, self-correction, no response due to anomia or speech arrest, performance errors, and errors due to effects on muscles. Errors can also be classified as clear or unclear depending on the certainty of the effect of stimulation. This coding scheme has since been used for language mapping with rnTMS. Lioumis et al. [2] proposed a setup for this procedure that allows offline review of the individual's performance involving synchronous video and audio recording in combination with high-precision navigated magnetic stimulation. Stimulated brain areas are displayed on a volume-rendered magnetic resonance image of the patient's brain, making it possible to improve localization accuracy and detect subtle errors. Most of the existing data on language mapping with rnTMS prior to our study are based on studies of adults. Although children are increasingly being investigated using rnTMS for language mapping, to our knowledge, no study has been published using the same procedure for children except for a study including four teenagers 15–18 years old and five adults [12].

https://doi.org/10.1016/j.yebeh.2018.05.036 1525-5050/© 2018 Elsevier Inc. All rights reserved.

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036

2

G. Rejnö-Habte Selassie et al. / Epilepsy & Behavior xxx (2018) xxx–xxx

In Sweden, children in need of neurosurgery are assessed for language mapping using a functional magnetic resonance imaging (fMRI) procedure [13]. Additionally, a dichotic listening test is routinely used for assessing language laterality in our center. However, assessing language with fMRI requires the child to lie completely tranquil and motionless in the magnetic resonance (MR) machine for a fairly long time and also to be able to follow test paradigms. This can be hard to achieve for a young child, particularly a child with problems of attention and perseverance. This means that fMRI is not suitable for language mapping in many children in need of presurgical assessment as some are very young and/or have various kinds of cognitive disadvantages. Repetitive navigated transcranial magnetic stimulation offers a different and easier way for cortical mapping of language as the child can be seated without restraints, even on the parent's lap, and breaks can be allowed if needed. It may also be possible to get more detailed information about localization of language areas when using rnTMS as this procedure blocks activity in crucial cortical locations while fMRI highlights a wider area of the language network during an activity [14]. Furthermore, in patients with brain tumors, rnTMS may be more useful for preoperative language mapping than fMRI as brain tumors may affect the anatomy and vasculature and thereby the blood flow and oxygenation in the brain, making fMRI less reliable [3]. The choice of words and images in a picture-naming task is important for the individual's performance, particularly the picture quality, the picture context, the familiarity of the depicted object, the length and familiarity of the name of the object, and the task variation, all of which can affect the speed of naming [15,16]. It is also important that pictures are familiar from a cultural point of view; otherwise, they may depict objects that are of a local character and therefore unfamiliar to the spectator. It is known that children have slower and less automatic language processes compared with adults [17], and it is likely that there is an age-dependent difference among children [13]. One proposal is that a starting phrase should be used with rnTMS as a way of sorting out whether the stimulation affects word retrieval or motor ability [4]. However, a starting phrase may decrease the naming speed and add to the difficulties with attention and perseverance, particularly in younger children. The setup for language mapping with rnTMS used in Sweden in adults has not previously been adjusted for use in children, so we conducted a methodological study to this effect. We evaluated the pictures used for testing of object naming and the preferred speed for naming so that the baseline testing could be shortened and the stress of testing could be reduced. The questions we asked were as follows: which pictures and which naming speed would be suitable to use for language mapping with rnTMS in children and in which children can these be used? The study resulted in a proposed method for use in children. We then studied the proposed method for preoperative language mapping with rnTMS among five children with difficult-to-treat epilepsy in a clinical setting. We analyzed both the results of the assessment and the individual adjustments made during the assessment with the aim of evaluating whether the method can be reliably used in children of different ages and abilities. We asked the following questions. When conducting an expressive language mapping using rnTMS in children with epilepsy, which naming errors will be found? Can language laterality or even localization of expressive language function be established? What individual adjustments need to be made? How do these children with epilepsy generally perform during the assessment? 2. Methodological study 2.1. Participants Forty healthy children and 10 children with speech and language disorder were assessed for object naming without transcranial magnetic stimulation. All tests were performed in Swedish. The healthy children were recruited by convenience and divided into four age groups: 10

children aged 3.10–6.10 years (M = 5.3 years), 13 children aged 7.5– 9.7 years (M = 8.6 years), 9 children aged 10.2–12.1 years (M = 11.5 years), and 8 children aged 13.1–15.2 years (M = 14.0 years). The majority (30/40) came from families with higher (academic) education. One child was bilingual. The children with speech and language disorder were divided into three age groups: 3 children aged 4–6 years (M = 5.6 years), 4 children aged 7–9 years (M = 8.3 years), and 3 children aged 10–12 years (M = 10.11 years). They were recruited from a clinical setting and displayed speech and language disorders of varying quality and magnitude such as difficulty with word retrieval, pronunciation, pragmatics, or stuttering, and, in one case, attention disorder and epilepsy. Two were bilingual. None of these participants were included in the later study with rnTMS. 2.2. Materials and methods The slide show was based on pictures by Snoodgrass and Vanderwart [18] and consisted of 150 black and white drawings of familiar objects (this is the setup distributed with our chosen rnTMS system; eXimia 3.2.1, Nexstim Ltd., Helsinki). The pictures were distinct in appearance and easy for a Swedish viewer to identify (Fig. 1). We wanted to form a series of pictures that would suit both younger and older children and those with language delay or disorder, so in a pilot study, all 150 pictures were shown to four children aged 4.6–7 years with speech and language delay. Of these, 31 objects which were unfamiliar to the children were discarded to form a shorter version of 119 pictures. This shorter version would put less demand on perseverance and attention than the longer one as the time for baseline testing would be shorter and fewer words would be unfamiliar to the children. In the study, the full version of 150 pictures was presented to children over 10 years of age (and to two 9-year-olds with extraordinarily good verbal ability), and the shorter version was shown to the children younger than 10. The pictures were displayed on a computer using a PowerPoint slide show with an initial interpicture interval (IPI) of 5 s, and the children were asked to name the pictures. Following this, the best IPI for each child was assessed. After watching the whole slide show, the child was asked if the speed was too fast, too slow, or just right. They were then shown a shorter series of the pictures with different IPIs (chosen from 3, 4, and 6 s as appropriate) in order to home in on their preferred IPI. The children also identified the preferred IPI when a starting phrase (“I say…”) was later added. Permission to perform the study was given by the local ethics committee, and written informed consent for participation was received from all patients and/or their parents. 2.3. Results Pictures that were unfamiliar or incorrectly named by 20% of the healthy children in each age group were discarded, as were six additional pictures which the children with language disorder found difficult to identify. This resulted in two new slide shows consisting of 123 pictures for children 10 years of age and above and 84 pictures for children below 10 years of age (Appendices 1 and 2). Our study revealed that children tended to answer with a word that was in line with their personal experience rather than using a classification word as an adult would usually do. For example, if shown a dog with long fur, an adult would generally say “dog” while a child might say “wolf” if they had never seen a dog of that type, or “Alaskan husky” if they knew such a dog. The children who participated were very eager to perform well as they normally would be at school when confronted with a task, so they frequently responded by providing synonyms in addition to their initial answer not only on the second viewing (as they saw the pictures twice) but often even on the first viewing. Concerning the preferred IPI, there were both an age-dependent difference and large individual differences despite age. The average preferred IPI speed was 4.5 s among children aged 4–6 years (6 of 10

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036

G. Rejnö-Habte Selassie et al. / Epilepsy & Behavior xxx (2018) xxx–xxx

3

Fig. 1. Four examples of the drawings used as stimuli: airplane, baby, banana, and book.

preferred 5 s), 4 s among children aged 7–9 years, 4 s among children aged 10–12 years, and 3.9 s in children aged 13–15 years. In all age groups, the preferred IPI ranged from 3 to 5 s. Children who requested a shorter IPI produced more naming errors, and this was particularly evident among the 7-year-olds. The children of the youngest age group had poor attention span. A starting phrase was refused by all children aged 4–6 years and some of the 7–9-year-olds. The children who used a starting phrase often preferred a 1 s longer IPI than when testing without this phrase while a few accepted the same speed. Children with language delay preferred on average a 5-s IPI without a starting phrase (range: 4–6 s) and were less willing to use the starting phrase before naming the object. 2.4. Conclusions and suggestions In our modified slide shows, we have adjusted the original procedure to suit different ages and be in line with contemporary culture and shortened the baseline testing by minimizing the number of pictures. We also propose that the IPI should be adjusted to 5 s for children aged 4–6 years and 4 s for children aged 7–15 years when a starting phrase is not used. For children with a known dysfunction of word retrieval, other language disorders, multilingualism, oral motor dysfunction, or stuttering, a longer IPI should be considered. However, individual adjustments should be made in all cases. A starting phrase should be avoided for children below 10 years of age and children with dysfunction of language and oral motor ability. When a starting phrase is used, the IPI might have to be increased by 1 s as compared with testing without this phrase. Furthermore, as children tend to use more synonyms than adults, when a child responds with a synonym during a clinical rnTMS session, it is not recommended to regard this as the result of the stimulation. 3. Clinical report on five children with epilepsy 3.1. Participants The clinical report included five children due to undergo epilepsy surgery: three girls and two boys aged between 8.4 and 12.11 years. They all had difficult-to-treat focal epilepsy with the following etiology: dysembryoplastic neuroepithelial tumors (DNET) in three (nos. 1, 2, and 4), focal cortical dysplasia (FCD) in one (no. 3), and sclerosis five years after operation and treatment of a primitive neuroectodermal tumor (PNET) in one (no. 5). One (no. 5) also had continuous spike– wave seizures during sleep. Duration of epilepsy varied from one to 11 years. Seizure frequency was daily in two (nos. 3 and 5), weekly in two (nos. 1 and 2), and monthly in one (no. 4). All participants were on various antiepileptic drug treatments (Table 1).

3.3. Methods Prior to the rnTMS language mapping, the children were assessed with neuropsychological tests for cognitive level, working memory, processing speed, visual reaction time, and general laterality [19–21]. Speech and language abilities were assessed regarding receptive grammar, expressive and/or receptive vocabulary, naming speed, and articulatory ability [22–26] (Table 1). Ear preference was assessed using the dichotic listening test of syllables with altering consonants as an indicator of language laterality [27]. The above-described procedures proposed by Lioumis and Corina for testing and evaluation, respectively, were followed. To individualize stimulation intensity, resting motor threshold (RMT) for the abductor pollicis brevis muscle (right hand in four children, left hand in one [no. 4]) was assessed as follows. After optimizing the location, angle, and direction of stimulation, the threshold was defined using the paradigm of the equipment (based on work by Awiszus [28]). The unit used for RMT and for the testing stimuli was the electrical field strength (V/ m) at the point of stimulation as calculated by the stimulation equipment. The children above 10 years of age (except child no. 5 because of poor perseverance) were presented with the long version of the picture show, and children below 10 years were presented with the short version. The time between picture and stimulation (picture trigger interval [PTI]) was 0 s in all children. In terms of starting phrase and IPI, the guidelines suggested in Section 2.4 were followed, but individual adjustments were made to each child's competence and comfort time. Picture display time was 2 s for all participants except child no. 4 where it was 2.6 s. The stimulation parameters — number of pulses, frequency (Hz), and intensity measured in electrical field strength (V/m) — were all individually adjusted. Stimulation was started using the individually defined RMT intensity. Intensity in terms of percentage of motor threshold varied during the investigation as the field strength in V/m varied in different sites, and the intensity was adjusted to maintain the same field strength. For all but one patient (no. 4), stimulus intensity had to be reduced because of perceived discomfort (Table 2). Both hemispheres were stimulated in four patients and only the left in the remaining one (no. 2) (Table 2). Total assessment time including preparations, establishment of RMT, and breaks ranged from 1 h 40 min to 2 h 15 min (Table 4). When analyzing the responses, errors were classified as “clear” when they could definitely not be attributed to other causes (e.g., poor attention, discomfort, or poor general articulation and language ability) and “unclear” when there was a suspicion of such a cause.

3.2. Materials 3.4. Results The two picture slide shows described above were shown on a monitor using a system which combined synchronous video and audio recording with rnTMS (eXimia NBS version 4.3.1, Nexstim Ltd., Helsinki, Finland).

3.4.1. Cognitive and language assessments A detailed report on patient characteristics is given in Table 1. For cognitive factors such as intelligence quotient (IQ), working memory,

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036

4

Patient no.

Sex Age

Etiology

Duration of epilepsy

EEG Seizure Antiepileptic characteristics frequency drugs

Cognitive level

1

M

8.4

DNET

1 year

Focal

Weekly

OXC

2

F

9.11

DNET

4 years

Focal

Weekly

VPA, LEV

Right hand and foot Slow Full-scale IQ: average Performance: IQ average Working memory: average Processing speed: average Right hand and Normal Full-scale IQ: average foot, left eye Verbal IQ: average Performance IQ: average Working memory: average Processing speed: average

3

M

11.6

FCD

7 years

Focal

Daily

VPA, LTG

4

F

12.11 DNET

11 years

Focal

Monthly

LTG

5

F

11.4

8.4 years

Focal and CSWS

Daily

VPA, LEV, TPM, CLOB

Sclerosis five years postsurgery, treatment of PNET

Laterality

Full-scale IQ: average Verbal IQ: average Performance IQ: low average Working memory: average Processing speed: low average Left hand and eye, Full-scale IQ: average right foot Verbal IQ: average Performance IQ: average Working memory: average Processing speed: average

Full-scale IQ: low average Verbal IQ: low Performance IQ: average Working memory: low Processing speed: average

Left hand and foot, right eye

Visual reaction time

Slow

Normal

Normal

Language level

Receptive grammar: within normal range Receptive vocabulary: within normal range Rapid naming: normal speed, some naming errors Articulation normal except for phonemes/l, r/ Receptive grammar: within normal range Receptive vocabulary: below age norms Rapid naming: slower and more naming errors than age norms Articulation: indistinct but normal in spontaneous speech Receptive grammar: within normal range Expressive vocabulary: above age norms Rapid naming: normal speed, some naming errors Articulation: normal

Receptive grammar: above age norms Receptive vocabulary: above age norms Expressive vocabulary: above age norms Rapid naming: normal speed and number of naming errors Articulation: inarticulate and slow speech, coarticulation difficulties Receptive grammar: below age norms Receptive vocabulary: below age norms Rapid naming: very slow and more naming errors than age norms Articulation: inarticulate and slow speech, phonological instability

No. = number, M = male, F = female, DNET = dysembryoplastic neuroepithelial tumor, FCD = focal cortical dysplasia, PNET = primitive neuroectodermal tumor, CSWS = continuous spikes during slow-wave sleep, OXC = oxcarbamazepine, VPA = valproate, LEV = levetiracetam, LTG = lamotrigine, TPM = topiramate, CLOB = clobazam, IQ = intelligence quotient, ms = milliseconds.

G. Rejnö-Habte Selassie et al. / Epilepsy & Behavior xxx (2018) xxx–xxx

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036

Table 1 Participant characteristics.

G. Rejnö-Habte Selassie et al. / Epilepsy & Behavior xxx (2018) xxx–xxx

5

Table 2 Methodological procedures used with rnTMS. Patient no.

RMT SO %

RMT V/m

Slide show version

IPI

Starting phrase

DT

No. of pulses

Frequency

Intensity % of RMT

Stimulated areas

No. of stimulations

1 2 3

60 91 55

130 165 160

Short Short Long

4s 4s 4s

No Yes Yes

2.0 s 2.0 s 2.0 s

10 10 10

5 Hz 5 Hz 5 Hz

67–83 49–82 45–100

Left + right Left Left + right

4 5

38 55

98 100

Long Short

4s 5s

Yes Yes

2.6 s 2.0 s

13 14

5 Hz 7 Hz

100–124 64–82

Left + right Left + right

289 246 254 left 143 right 316 179

No. = number, SO = stimulator output, V/m = volts per meter, IPI = interpicture interval, s = second, DT = picture display time, Hz = Hertz, RMT = resting motor threshold.

and processing speed, scores between 85 and 115 are classified as average, those between 70 and 84 as low average, and those lower than 70 as low. The majority had average scores, a few had low average scores, and one participant (no. 5) had a low verbal IQ score. Visual reaction times of above 400 milliseconds (ms) are regarded as slow and those below 400 ms as normal; two participants (nos. 1 and 3) were classified as slow and the remaining three as normal. General laterality was not measured in participant 3 and was not clear in the others. Grammar and vocabulary were below age norms in one (no. 5). The speed of rapid naming was normal in three (nos. 1, 3, and 4) and below normal in two (nos. 2 and 5) while number of naming errors during rapid naming was only normal in one participant (no. 4). Articulation was slow and indistinct in three participants (nos. 2, 4, and 5) (Table 1). 3.4.2. Object naming errors from rnTMS Naming errors judged as clear were classified as no response, semantic error, phonological error, self-correction, and response delay. The most common error was response delay followed by no response, semantic error, phonological error, and self-correction (Table 3). Thus, all the common types of error described in studies on adults were also found in these children. 3.4.3. Language laterality from rnTMS and the dichotic listening test Table 3 presents the types of naming error assessed as clear for each hemisphere. Participant 2 was only stimulated over the left hemisphere, so no conclusions about laterality can be drawn from that investigation. The dichotic listening test indicated that participants 1 and 4 had a left ear preference, 2 and 3 a right ear preference, and 5 no preference (Table 3). Ear preference according to the dichotic listening test is contralateral to the language-dominant side (Table 3). 3.4.4. Localization from rnTMS The concept of reproducibility for this kind of testing of language function localization is controversial even when performed in adults. Given the practical difficulties when testing children, no attempt was made to approach this subject in the present study. An example of the results obtained is given in Fig. 2. 3.4.5. Individual adjustments 3.4.5.1. Breaks. The children were given both long and short breaks during the assessment. Breaks longer than 4 min were classified as long and those shorter than 4 min as short. The latter were all between 0:21 and 2:44 min. Several short breaks were given because of discomfort from

simultaneous stimulation of the jaw muscles, particularly in participants 1, 2, and 3 and, on a few occasions, also in participant 5. In other instances, breaks were given because of fatigue, poor perseverance or concentration, or technical issues. Participant 5 had very limited attention span and increased fatigability and hence required many short breaks. Long breaks were given when the stimulation shifted to the opposite side and for reasons of fatigue or poor perseverance (Table 4). 3.4.5.2. Reduction of intensity. The participants experienced some degree of discomfort from simultaneous stimulation of the jaw muscles, which led to reduction of intensity in participants 1, 2, 3, and 5. The intensity was reduced stepwise by 2–3 percentile units each time until the stimuli were perceived as comfortable (Tables 2 and 4). 3.5. Performance and adverse events No seizures were detected during the assessments, and all participants cooperated well. The youngest participant (no. 1) was somewhat restless, but interest in the technical aspects of the procedure enhanced his perseverance. Because of discomfort which puts strain on perseverance, participant 2 was only stimulated in the left hemisphere. Participant 5 had a very short attention span, leading to several breaks and limitations of the assessment. 4. Discussion The aim of this study was to find the optimum method of presenting a picture slide show for object naming to children of different ages and performance capacities in order to establish a protocol to be used in a preoperative cortical mapping of expressive language with rnTMS in children. The parameters studied were the choice of pictures and the IPI. The proposed protocol as well as the need for individual adjustments of stimulus parameters was subsequently tested by performing rnTMS for language mapping in five children with epilepsy who were due to undergo epilepsy surgery. When choosing a picture slide show to be used for object naming in children, it is important that both the words and the depicted objects are familiar to the child. The two slide shows consisted of words and pictures that were familiar to Swedish children and were adapted for two age groups: above and below ten years of age, respectively. The original slide show used for adults contained some pictures that were old-fashioned or not age-appropriate, and the series was long, which could lead to fatigue and frustration over poor performance. After establishing a baseline of competence in an rnTMS assessment, pictures which are

Table 3 Naming errors induced by rnTMS (number per hemisphere) and assessed as clear, and ear preference assessed with the dichotic listening test. Patient no.

No response

Semantic error

Phonological error

Self-correction

Response delay

Total

Ear preference

1 2 3 4 5

3 left/3 right

5 left/1 right 2 left 1 left/3 right 2 right 1 right

2 left/1 right 1 left 2 right 1 right –

– 1 left 2 right 2 right 1 left/1 right

14 left/4 right 6 left 8 left/5 right 3 left/5 right 2 left/4 right

24 left/9 right 10 left 18 left/36 right 7 left/13 right 3 left/7 right

Left Right Right Left No preference

9 left/24 right 4 left/3 right 1 right

Patient no. = patient number.

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036

6

G. Rejnö-Habte Selassie et al. / Epilepsy & Behavior xxx (2018) xxx–xxx

Fig. 2. Results from the language testing in one of the children (no. 1). Both hemispheres were examined. Markers show the localizations of the stimuli applied, and their colors denote the effects of stimulation. Red: response delay and phonological error, yellow: semantic error, white: anomia, and gray: no effect.

incorrectly named are discarded from further use. This produces a shorter series of pictures for use among children, thus minimizing the number of naming errors and resulting in less fatigue and frustration. It is also important that pictures are culturally adapted, so the choice of words/pictures for our population may not fit all groups of children. The preferred IPI was longer in children than in adults. This is in accordance with other studies showing that children have less automatic language processes compared with adults [17]. An age-dependent effect among children resulted in our recommending different IPIs for different ages. Aside from one short report [29], no earlier studies using rnTMS in young children were available at the time when this study was conducted. However, our clinical report revealed other differences between the children and adults (who are routinely assessed preoperatively with rnTMS in centers using this technique). Total assessment time was longer for all the children than for adults. Children often need several breaks as perseverance is poorer in younger participants. Giving children plenty of total assessment time would therefore make the results more stable. Slow visual reaction time, which was found in two participants, may also determine the preferred IPI and picture display time. General cognitive ability, attention, working memory, and processing speed may interfere with the results of this process, particularly among children with epilepsy and other neurodevelopmental disorders, and fluctuating abilities may exist even without stimulation [30]. Furthermore, children with a language competence below average, as found in one participant in the present study, may have slower language processing capacity because of difficulty with phonological working memory [31]. Thus, children should routinely be tested for cognitive, speech, and language competence prior to an rnTMS language mapping. This is also important when it comes to correctly evaluating possible naming errors as particular difficulties in this domain, including word retrieval, may exist for children with epilepsy or other neurodevelopmental conditions [32]. The most common naming error was response delay. This may have been an effect of the stimulation, but according to Corina et al. [11] and Lioumis et al. [2], there is uncertainty over how to judge response

Table 4 Individual adjustments made during rnTMS assessments. Participant no.

Long breaks

Short breaks

Total assessment time

Reduced intensity

1 2 3 4 5

3 3 3 1 2

13 11 19 7 24

1 h 50 min 2 h 15 min 2 h 10 min 1 h 40 min 2h

Yes Yes Yes

No. = number, min = minutes, h = hour.

Yes

delays. Delays may be due to factors other than stimulation. When evaluating these response delays, the child's general processing speed, vocabulary, naming speed, articulation, and attentional capacity have to be considered. Some of the children in the clinical study had slow word retrieval, and the majority had slow and indistinct articulation. However, since each child is their own control (comparing with baseline testing), response delays may be relevant to stimulation effects and are judged as errors in several rnTMS registrations. The most reliable effect of stimulation is considered to be no response, which was the second most common error in our study. Semantic or phonological errors should also be seen as clear indications of a stimulation effect as they are examples of disrupted expressive language. The question of language laterality or localization of expressive language is not simple. No attempt was made to approach the subject of language function localization testing in the present study as it is controversial even when performed in adults. The clear naming errors induced by the rnTMS showed that hemispheric dominance of expressive language varied between the participants: more to the left in participant 1, to the right in participant 3, and without a clear side difference in participants 4 and 5. However, bilateral stimulation effects were seen in all cases. The rnTMS assessment and the dichotic listening test showed results consistent with each other in participants 4 and 5 but were inconsistent in participants 1 and 3. Several parts of the language network are involved in object naming. According to Indefrey [15], the word is retrieved from the lemma store, which involves the left middle temporal gyrus, and then the phonological code for the word form is retrieved by activation of the left posterior superior temporal gyrus. Before word production takes place, the phonological encoding of the word and the motor planning are processed in the left inferior frontal gyrus (Broca's area). In a dichotic listening test, a phonological judgment of similarity between syllables is made with no involvement of word retrieval. For these reasons, object naming and dichotic listening may not produce the same results, but the two methods may be complementary. The younger the child is, the more likely they are to also show right hemispheric involvement as the child first learns the word as a whole without the phonological analysis [33]. Interestingly, in participant 1, ear preference had shifted to a right ear dominance postsurgery, which may indicate that the rnTMS assessment of laterality was the more accurate one or may be due to fluctuating laterality as a result of epileptiform activity. Furthermore, none of our participants showed a clear laterality when evaluating dominant ear, hand, foot, and eye, indicating a less developed laterality as a whole. Our participants all had epilepsy, and epilepsy in a maturing brain may alter the language network [34]. All of this means that when a clear language laterality is not found using the rnTMS technique, it does not mean that the method is unreliable. Furthermore, as language is supported by a wide cerebral network, it is improbable that one could localize a specific site responsible for expressive language. Instead, the main contribution of rnTMS testing for

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036

G. Rejnö-Habte Selassie et al. / Epilepsy & Behavior xxx (2018) xxx–xxx

cerebral mapping of expressive language may be the assessment of cortical function around a potential surgical site to minimize the risk of a negative effect of the surgery. The fact that all our participants had epilepsy demonstrates that it is possible to assess laterality of expressive language and cortical function around defined cortical areas in this patient group. Repetitive navigated transcranial magnetic stimulation may be a particularly suitable procedure as it allows the individualized approach that these children generally need during the assessment. The focus of this study was to construct a paradigm for object naming with rnTMS that would be well-adjusted for use in young children, but a validation of its usefulness compared with results from other methods of language mapping needs to be made. Such a validation of results from preoperative language mapping with rnTMS in patients with epilepsy was performed in the study by Babajani-Feremi et al. [12] and in a more recently published study by Lehtinen et al. of children with epilepsy [35]. These studies compared the results from rnTMS with results from fMRI, direct cortical stimulation, and, in Babajani-Feremi et al.'s study, high-gamma electrocorticography. There was good agreement between the methods tested, indicating that our findings may have practical relevance which gives support to our method. The pediatric population in the study by Lehtinen et al. contained five children 9–12 years of age and nine teenagers. This young age group corresponds to our population of 8–12 year olds while a group of teenagers might have more similarities with adults when it comes to language proficiency and ability to cooperate during the assessment. Thus, the present study adds information concerning the applicability of this method even in young children. 4.1. Limitations Children may experience more discomfort from the stimulation than adults. This may partly be due to high stimulus intensity since the RMT is usually higher in children than in adults. We therefore reduced the intensity in this study thus increasing the risk of producing nonefficient stimulation. 4.2. Conclusion We have adjusted a picture slide show for adults for use with rnTMS in language mapping in children of different ages and evaluated its usefulness in five children with epilepsy. We found that this procedure is feasible, well-functioning, and suited for making individual adjustments within the testing frame, which makes it preferable for children where fMRI cannot be used for language mapping. The children tested had difficult-to-treat epilepsy and were due to undergo epilepsy surgery. They also had neurodevelopmental comorbidities that made the testing challenging. Language lateralization was compared with results from cognitive assessments of general laterality and language laterality via a dichotic listening test and found compatible in half of the cases. However, general laterality was unclear in several of the participants. Assessment of language localization was not attempted in this investigation. In a small group, we have shown that it is possible to use this procedure in children with difficult-to-treat epilepsy at least from eight years of age. However, for a more detailed evaluation, a larger study group of young children would need to be assessed and results to be compared with those of other methods of language mapping. Acknowledgments We would like to thank all the participating children and their parents. We are also grateful to Mikael Elam, Åsa Nordberg, Elsa Trenning, and the epilepsy surgery team at Queen Silvia Children's Hospital. This work was financially supported by Föreningen Margarethahemmet, which had no other involvement in the project.

7

Declaration of interests MT owns shares in the company (Nexstim Ltd.), producing the stimulation equipment used in the study. Other authors have none to declare. Appendix 1. List of 84 words for children below 10 years of age

Airplane Apple Arrow Axe Baby Backpack Ball Balloon Banana Basket Bench Bicycle Binoculars Bird Boat Bone Book Bowl Bread Bucket Butterfly Cake Camera Candle Car Carrot Cat Chair

Cheese Clock Dog Dress Drum Fence Fire Flower Foot Football Fork Frog Glass Glasses Guitar Hamburger Hammer Hand Hat Heart Horse House Ladder Leaf Lips Lizard Microphone Mouse

Mushroom Needle Onion Pencil Pen Necklace Piano Pig Present Rake Robot Saw Scissors Shark Spade Skateboard Skis Snake Sock Spoon Stairs Suitcase Sword Toothbrush Torch Trousers Umbrella Vacuum cleaner

Appendix 2. List of 123 words for children 10 years of age and above

Airplane Apple Arrow Axe Baby Backpack Ball Balloon Banana Barrel Basket Bench Bicycle Binoculars Bird Biscuit Boat Bomb Bone Book Bottle Bowl Box Bread Bucket Butter Butterfly Cake

Camera Candle Car Carrot Cat Chair Cheese Clock Corn Coin Dog Doll Dress Drum Ear Egg Fan Feather Fens Fire Flower Foot Football Fork Frog Frying pan Funnel Glass

Glasses Globe Glove Guitar Hamburger Hammer Hand Hat Heart Helmet Horse House Jar Ladder Lamp Leaf Lips Lipstick Lizard Lock Microphone Mouse Mushroom Necklace Needle Onion Paintbrush Paperclip

Pen Pencil Phone Piano Pig Postbox Present Radio Rake Robot Saucepan Saw Scissors Screwdriver Shark Shoe Spade Skateboard Skis Snake Sock Spoon Stairs Stool Suitcase Sword Tap Tie

Tomato Tooth Toothbrush Torch Trousers Trumpet Umbrella Vacuum cleaner Wheelbarrow Wheelchair Yoyo

References [1] Vitikainen A-M, Lioumis P, Pateau R, Salli E, Komssi S, Metsähonkala L, et al. Combined use of non-invasive techniques for improved functional localization for a selected group of epilepsy surgery candidates. Neuroimage 2009;45:342–8.

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036

8

G. Rejnö-Habte Selassie et al. / Epilepsy & Behavior xxx (2018) xxx–xxx

[2] Lioumis P, Zhdanov A, Mäkelä N, Lehtinen H, Wilenius J, Neuvonen T, et al. A novel approach for documenting naming errors induced by navigated transcranial magnetic stimulation. J Neurosci Methods 2012;204:349–54. [3] Sollman N, Picht T, Mäkelä J, Meyer B, Ringel F, Krieg S. Navigated transcranial magnetic stimulation for preoperative language mapping in a patient with a left frontoopercular glioblastoma. Case report. J Neurosurg 2013;118(1):175–9. [4] Tarapore P, Findlay A, Honma S, Mizuiri D, Houde J, Berger MS, et al. Language mapping with navigated repetitive TMS: proof of technique and validation. Neuroimage 2013;82:260–72. [5] Krieg S, Sollman N, Hauck T, Ille S, Foerschler A, Meyer B, et al. Functional language shift to the right hemisphere in patients with language-eloquent brain tumours. PLoS One 2013;8(9):1–9. [6] Picht T, Krieg SM, Sollman N, Rösler J, Niraula B, Neuvonen T, et al. A comparison of language mapping by preoperative navigated transcranial magnetic stimulation and direct cortical stimulation during awake surgery. Neurosurgery 2013;72(5):808–19. [7] Säisänen L, Könönen M, Julkunen P, Määttä S, Vanninen R, Immonen A, et al. Non-invasive preoperative localization of primary motor cortex in epilepsy surgery by navigated transcranial magnetic stimulation. Epilepsy Res 2010;92:134–44. [8] Thordstein M, Hallböök T, Lundgren J, van Westen D, Elam M. Transfer of cortical motor representation after a perinatal ischemic cerebral insult. Pediatr Neurol 2011;44:131–4. https://doi.org/10.1016/j.pediatrneurol.2010.08.017. [9] Petrovich Brennan NM, Wahlen S, de Morales Branco D, O'Shea JP, Norton IH, Golby AJ. Object naming is a more sensitive measure of speech localization than number counting: converging evidence from direct cortical stimulation and fMRI. Neuroimage 2007;37(Suppl.1):S100–8. [10] Hernandez-Pavon J, Mäkelä N, Lehtinen H, Lioumis P, Mäkelä J. Effects of navigated TMS on object and action naming. Front Hum Neurosci 2014. https://doi.org/10. 3389/fnhum.2014.00660. [11] Corina DP, Loudermilk BC, Detwiler L, Martin RF, Brinkley JF, Ojeman G. Analysis of naming errors during cortical stimulation mapping: implications for models of language representation. Brain Lang 2010;115:101–12. [12] Babajani-Feremi A, Narrayana S, Rezaye R, Choudhri AF, Fulton SP, Boop FA, et al. Language mapping using high gamma electrocorticography, fMRI and TMS versus electrocortical stimulation. Clin Neurophysiol 2016;127:1822–36. [13] Norrelgen F, Lilja A, Ingvar M, Gisselgård J, Fransson P. Language lateralization in children aged 10 to 11 years: a combined fMRI and dichotic listening study. PLoS One 2012;7(12):e51872 [1–11]. [14] Forster M-T, Hattingen E, Senft C, Gasset T, Seifert V, Szelényi A. Navigated transcranial magnetic stimulation and functional magnetic resonance imaging: advanced adjuncts in preoperative planning for central region tumours. Neurosurgery 2011; 68:1317–25. [15] Indefrey P. The spatial and temporal signatures of word production components: a critical update. Front Psych 2011;2(255). [16] Cibelli ES, Leonard MK, Johnson K, Chang EF. The influence of lexical statistics on temporal lobe cortical dynamics during spoken word listening. Brain Lang 2015; 147:66–75.

[17] Brauer J, Neumann J, Friederici AD. Temporal dynamics of perisylvian activation during language processing in children and adults. Neuroimage 2008;41(4):1484–92. [18] Snoodgrass JG, Vanderwart M. A standardized set of 260 pictures: norms for name agreement, image agreement, familiarity, and visual complexity. J Exp Psychol Hum Learn 1980;6(2):174–215. [19] Wechsler D. Wechsler Intelligence Scale for Children. 4th ed. Stockholm: Harcourt Assessment; 2007 (Swedish version). [20] Reitan RM, Davidson LA. Clinical neuropsychology: current status and applications. New York: Hemisphere; 1974. [21] Rugland A-L, Henriksen O, Bjornaes H. Computer-assisted neuropsychological assessment in patients with epilepsy. In: Dodson WE, Kinsbourne M, Hiltbrunner B, editors. The assessment of cognitive function in epilepsy. New York: Dermos Publications; 1991. p. 109–25. [22] Bishop D. Test for reception of grammar. Version 2 (Swedish version). London, Stockholm: The Psychological Corporation, Harcourt Assessment; 2003. [23] Dunn L, Dunn LI. PPVT-II, Peabody Picture Vocabulary Test. 3rd edition. SIH Läromedel; 2001 (Swedish version). [24] Kaplan E, Goodglass H, Weintraub S. Boston Naming Test. 2nd edition. Austin, Texas: PRO-ED Inc.; 2001. [25] Eklund H. In: Schöndin K, Svensson L, editors. Rapid confrontation naming: Swedish assessment material for word retrieval. Normative data and evaluation. Göteborg, Sweden: Department of Speech and Language Pathology, Sahlgrenska Academy at University of Gothenburg; 1996. [26] Holmberg E, Sahlén B. Nelli: neurolinguistic assessment battery for language disordered children (Swedish). Malmö: Utbildningsproduktion; 1986. [27] Hugdahl K. Dichotic listening in the study of auditory laterality. In: Hugdahl K, Davidson RJ, editors. The asymmetrical brain. Cambridge, MA: MIT Press; 2003. p. 441–76. [28] Awiszus F. TMS and threshold hunting. Suppl Clin Neurophysiol 2003;56:13e23. [29] Krieg SM, editor. Navigated transcranial magnetic stimulation in neurosurgery. Springer International Publishing AG; 2017. [30] Aldenkamp AP, Reijs R, Van Mil S, Debeij-Van Hall M. Attention disorders in children with epilepsy. 26th IEC Proceedings. Epilepsia 2005;46(suppl. 6):6. [31] Montgomery J. Working memory and comprehension in children with specific language impairment: what we know so far. J Commun Disord 2003;36(3):221–31. [32] Rejnö-Habte Selassie G, Viggedal G, Olsson I, Jennische M. Speech, language and cognition in preschool children with epilepsy. Dev Med Child Neurol 2008;50:432–8. [33] Nora A, Karvonen L, Renvall H, Parviainen T, Kim J-Y, Service E, et al. Children show right-lateralized effects of spoken word-form learning. PLoS One 2017;12(2): e0171034. https://doi.org/10.1371/journal.pone.0171034. [34] Deonna T, Roulet-Perez E. Cognitive and behavioural disorders of epileptic origin in children. Mac Keith Press; 2005. [35] Lehtinen H, Mäkelä JP, Mäkelä T, Lioumis P, Metsähonkala L, Hokkanen L, et al. Language mapping with navigated transcranial magnetic stimulation in pediatric and adult patients undergoing epilepsy surgery: comparison with extraoperative direct cortical stimulation. Epilepsia Open 2018:1–12. https://doi.org/10.1002/epi4.12110.

Please cite this article as: Rejnö-Habte Selassie G, et al, Navigated transcranial magnetic stimulation for preoperative cortical mapping of expressive language in children: Development of a..., Epilepsy Behav (2018), https://doi.org/10.1016/j.yebeh.2018.05.036