Effect of Stimulation of Dorsolateral Prefrontal Cortex on Language Control

Effect of stimulation of dorsolateral prefrontal cortex on language production in healthy bilinguals

Background:

Language relies on two main neurocognitive components; a dedicated language network depending on perisylvian structures, and a domain-general control-executive and working memory system relying on prefrontal, particularly left cingulum and dorso-lateral prefrontal cortex (DLPFC) networks. The executive system is known to participate in language control in bilinguals, but also plays a role in many intra-languages processing, for mother language (L1) and even more in second language production (L2).

Anodal-tDCS appears to improve cognitive functions (e.g. working memory, complex verbal problem-solving, inhibition) in adults. Despite the fact that the contribution of executive function in language processing in monolingual and multilingual speakers is crucial, it has not received considerable attention.

Aim:

This study aims to demonstrate the impact of the modulation of executive functions, particularly subserved by DLPFC, on language production in healthy bilingual. The investigators propose that such impact will be particularly present in L2.

The experimental task will test the impact of prefrontal stimulation by tDCS (Anodal compared to Sham) over left DLPFC of healthy adult bilingual subjects during two tasks relying on lexical strategies: a picture naming task in L1 and L2 and a word translation task. The EEG correlates of these tasks performed following tDCS will be analysed, which allows us to investigate precisely the spatiotemporal brain dynamics underlying the behavioural modifications induced by the modulation of left frontal areas. The hypotheses are:

1. A-tDCS-induced left DLPFC activation will increase performance particularly in L2 naming and in the translation task versus baseline performances (sham tDCS).
2. A-tDCS is expected to modify the cortical activity related to the language tasks. During naming, EEG analyses will show a main effect of stimulation and an interaction between the factors Language (L1 or L2) and Stimulation (sham, anodal) at the topographic level, showing an effect of prefrontal areas larger in L2 than L1. The latency of the ERP modulation by tDCS in L2 will allow us to target the dynamics and the language planning processes underlying improved naming performance.

Methods:

The experimental task will include healthy L1-L2 unbalanced subjects. The investigators will analyse the influence of the DLPFC stimulation on naming and translation ability using tDCS as well as their EEG correlates in L1 and L2. The prediction is that left prefrontal Anodal-stimulation will increase language performances particularly in L2. EEG analyses will unravel the spatio-temporal dynamics of executive processes involved in the two tasks.

EEG EEG is a direct non-invasive technique for recording of the electrical activity produced by neurons over the scalp with a high temporal resolution. EEG recording is obtained by placing electrodes on the scalp. The event-related potentials will be processed/analysed offline.

tDCS TDCS is an affordable and safe method of brain stimulation which can be broadly used in general population. In tDCS, the cerebral cortex is stimulated through a weak constant electric current (1-2 mA) through two electrodes (35 cm2 and 70cm2, an "active" electrode which targets a special region of brain and a "return" electrode). This weak current can induce focal changes of cortical excitability that lasts beyond the period of stimulation. In general, Anodal stimulation (A-tDCS) is supposed to depolarize neurons leading to an increase in excitability, whereas Cathodal stimulation (C-tDCS) has the opposite effect compared to sham. Sham tDCS (S-tDCS) refers to a control condition in which the subject will receive a brief current but then the stimulator remains off for the rest of the stimulation time. In this way, the individuals receiving the S-tDCS are not aware that they are not receiving prolonged stimulation.The mechanisms of tDCS effect are classified into synaptic (changes by altering the strength of synaptic transmission) and non-synaptic (shifts in resting membrane potential of pre and post-synaptic neurons).

Language proficiency assessment To assess language immersion, participants will be asked for the age of acquisition, how long they lived in a region where French and English language was spoken, the language spoken in family, in childhood, in current daily life, and if the language was acquired in or out of school. In the self-evaluation part, participants will indicate in an analogue scale from 1 to 100% how they estimate their reading, speaking, comprehension and writing skills. Moreover, a sub-test from the computer-based DIALANG language diagnosis system will be used to evaluate reading performance.

Procedure This study consists of two sessions, each one week apart, in a counterbalanced order: 1. Anodal A-tDCS, 2. Sham-tDCS. In both sessions, the same evaluations/tasks will be used.

• A-tDCS intervention: The active anodal electrode is placed over DLPFC and the return electrode over contralateral supraorbital area. A 1.5mA direct current will flows between the electrodes for 20 minutes. To decrease current-induced injuries, at the beginning of stimulation, the current reaches from 0 to 1.5mA through 30 seconds (fade-in), and at the end will decrease from 1.5mA to 0 in 15 seconds (fade-out).
• Sham-tDCS intervention: The parameters mimics the A-tDCS (i.e. electrode placement, fade-in and current intensity), except the duration of the stimulation: the current will reach to 1.5mA in 30 seconds (fade-in) and lasts only for 30 seconds (to give the initial sensation of stimulation) then decreases to 0.

The two language contexts will be presented within each of the two sessions in a counterbalanced order across subjects. At the beginning of each language context session, a short text written in the corresponding language will be showed to the subject in order to activate the language of interest. Participants will be seated in an electrically shielded and sound attenuated booth in front of a LCD screen. Stimulus delivery and response recording are controlled using E-Prime 2.0. The stimuli will be presented in the center of the screen.

Stimuli and tasks Verbal and non-verbal fluency tasks, without EEG recording At minute 15/20 of tDCS, a verbal fluency task in L1 will be done (the subject will be asked to name as many words as he/she can which starts with a certain letter during 1 minute). At 16/20 minutes, a non-verbal fluency task (5 points test, i.e. a sheet of paper with 40 squares of fixed symmetrically arranged points (5 points) will be presented to the participant. The participant will be asked to produce as many different figures as possible with connecting the dots with at least one line connecting two dots during 2 minutes). And at minute 18/20, a verbal fluency in L2 will be done during 1 minute (according to Nitsche and Paulus the effect of tDCS reaches maximum at 4/20 minutes of stimulation).

Naming task in L1 and L2 during EEG measures The naming task will be done based on a corpus of 70 pictures, the corresponding words being matched in English and French on pertinent psycholinguistic factors. The stimuli consist of black and white line drawings (of 8.5 x 8.5cm each) representing manufactured objects (tools, furniture, clothes, kitchen objects, electric apparatus, vehicles etc.) or living objects, selected from the Snodgrass and Vanderwart image bank according to lexical and pre-lexical norms and databases. The naming task consists of 2 blocks, one in L1 and one in L2 (each block with 70 images) with a 2-minute pause between the two blocks. The order of blocks will be counterbalanced. In each trial, an image will be presented centrally (1000 ms) after a fixation cross of 300 ms. The subject then will have maximum 2000 ms to name the object.

Translation task to L1 and L2 during EEG measurement Then, they will performed an offline task consisting in translating during 8 minutes 100 words (5 letters length) from L1 to their second language (L2) or vice versa. Both senses of translation will be mixed randomly to increase working memory load. Translation corpus consists of 50 words in each language, matched across L1 and L2 on frequency, length in syllables and phonemes. Words will be presented on the screen and the subject will be asked to translate each word in the other language. The words will be presented 1500 ms on the screen after a fixation cross of 300ms. The subject will have a maximum 2500ms to do the translation.

Behavioural and EEG analyses Behavioural analysis A first behavioural analysis will compare naming and translation performances, as well as speech onset latencies in both languages.

Statistical analyses will be conducted according to 2X2 experimental designs with Language (L1; L2) and Stimulation (Anodal, Sham) as within-subject factors. The investigators main interest will be the interaction term between Language and Stimulation. The investigators expect the interaction to be driven by the left DLPFC stimulation with A-tDCS improving performance in L2 but not L1.

Event Related Potentials (ERP) analyses ERPs will be analyzed using a step-wise procedure that permits the time-wise, multifactorial statistical assessment of (1) the dynamic changes of scalp-recorded electric field configuration and their temporal segmentation into quasi-stable functional micro-states indexing modulations of the configuration of intracranial generators, (2) global electric field power indexing modulation of response strength of the intracranial generators, (3) intracranial distributed linear electrical source estimations. These analyses will be conducted for each time frame of the whole ERP time-period; however, specific time windows are planned to be sensitive to the investigators factors according to meta-analyses and the investigators previous results: lexical selection is thought to occurs between 170 and 270 ms after picture presentation, so there will be a first time window until 200 ms, a second time window between 200 and 300 ms, and a later time window (300-450 ms) corresponding to lexical-phonological encoding in previous studies. The main effect will be L1 and L2 patterns in term of amplitude, duration and sequences. Then, a second main effect will be the anodal tDCS / sham tDCS effect in an early time window before the vocalization will start. The investigators will expect an interaction between language and stimulation, due to the fact that left DLPFC anodal-tDCS will modulate L2 more than