Neural correlates of inner speech – INNERSPEECH

fMRI data collection under different conditions of inner speech : wilful inner speech, mind wandering, auditory verbal imagery with allocentric and egocentric perspectives.
iEEG data collection under different conditions of inner speech : wilful inner speech, mind wandering, auditory verbal imagery with allocentric and egocentric perspectives, inner dialogue
EMG data collection during induced verbal mental ruminations (excessive inner speech) and effect of orofacial relaxation on rumination
TMS-EMG-EEG data collection : in progress

The fMRI data collected during the project have allowed us to describe the neural networks associated with the production of inner speech with our own voice vs. someone else’s voice (imitation) as well as those associated with imaginary speech perception vs actual speech perception. Taken together with the introspective questionnaires, they also enable us to distinguish between the networks of wilful inner speech and verbal mind wandering.
The collection and analysis of iEEG data in implanted patients is still in progress.
The study on mental rumination has been developed thanks to collaborations with the Laboratoire Interuniversitaire de Psychologie in Grenoble, and the Psychopathology and Affective Neuroscience Lab in Ghent University (Belgium). Our EMG findings suggest that the lip motor system is involved in induced ruminations. They provide promising strategies to reduce rumination via orofacial relaxation or inhibition.
The analysis of the detailed introspective data collected in collaboration with the Monologuer Project (S. Smadja, Univ. Paris Diderot) is still under way.
A TMS experiment aiming at reducing induced rumination is in progress, in collaboration with Ghent University.

The study of mental rumination has enabled us to launch a collaboration with the LIP in Grenoble as well as with the Psychopathology and Affective Neuroscience Lab, in Ghent Univ. (Belgium).
The feedbacks we received from the people who participated in the InnerSpeech – Monologuer experiment (colaboration with S. Smadja, Université Paris Diderot) show that keeping a diary on inner speech moments and participating in guided interviews are beneficial (finding time for oneself and setting back). This beneficial and perhaps therapeutic outcome seems relevant and we are currently working on a project proposal with S. Smadja (who submitted an ERC grant proposal in relation to this).
This project opens outcomes for rumination remediation, via labial relaxation or inhibition. We are currently setting a TMS protocol, in collaboration with Ghent university, to try to reduce ruminations by inhibiting motor planning processes.
This project also opens new perspectives for the remediation or care of auditory verbal hallucinations. In this line, we have contacted Charles Fernyhough, from Durham University, who is in charge of the Hearing Voices project. We are considering future collaborations with his group in this domain. We have also been contacted by the PsyCHIC group from SCAlab in Lille, run by Renaud Jardri, who is an expert in the neural mechanisms of hallucination. We are also considering future collaborations with this group.
Finally, the InnerSpeech project opens long term perspectives on the beneficial role of inner speech in the development of self-awareness and cognitive performance. In this line, we will continue our scientific exchanges with Stéphanie Smadja from Université Paris Diderot and we hope to start a new collaboration with Agustin Vicente from Basque country University in San Sebastian.

We have proposed a neurocognitive modelling of inner speech, published in the book Inner Speech: New Voices, in press at OUP.
We have contributed to several volumes, including the Traité de Neurolinguistique edited by Pinto and Sato (2016), a special issue on embodied cognition of the journal Recherche sur la Philosophie et le Langage (in press), a special issue on « Discours et espace intérieurs » of the journal Epistémocritique (submitted).
Our EMG data on rumination have been published in the Biological Psychology journal and presented in several international conferences. The motor interference experience on rumination is under submission in Psychological Research.
Our fMRI data have been presented in several conferences.
We have been invited to give scientific talks in several European universities and have given talks to the general public.
We have participated in a television program broadcast on Radio Canada TV and in several French and Canadian radio shows. We have contributed to several articles in newspapers and magazines.
We have organised an international workshop on inner speech on 7 June 2016 as well as a colloquium (with international guest speakers), on the forms, functions and dysfunctions on inner language on 23 June 2017. We have also co-organised a Monologue-InnerSpeech workshop with S. Smadja on the topic of inner speech experiences in Paris, in February 2017.

Inner speech (IS) refers to the production of words in one’s mind. It has a positive influence on many cognitive tasks, such as memory, thinking, computation and self-awareness. It can be disrupted and have a more negative role, however. Excessive negative IS (mental rumination) in depression or anxiety disorder can interfere with cognitive performance; disruption of IS in schizophrenia may result in auditory verbal hallucinations (AVH), where one’s own verbal thoughts are perceived as external voices. The INNERSPEECH project will attempt to improve our understanding of IS by addressing 4 questions.
The first question – “WHAT” – relates to the nature of IS. Is IS closer to mental simulation of overt speech or to speech recall from memory traces (speech listening)? Which cerebral networks are at the origin of IS? What are their dynamics?
The second – “HOW” – concerns the way IS manifests itself. IS can be deliberate, when e.g. we mentally count objects, or more unconstrained, when our “mind wanders” during resting states. Is mind wandering neurally different from voluntary IS?
The third – “WHO” – deals with agency in IS. How do we know that the inner voice we hear is self-generated and not produced by someone else? We can hear our inner voice timbre and pitch. We can have imaginary dialogues with close relatives and we then hear their voices. How do we not confuse these inner voices with external voices?
The last – “WHAT PURPOSE” – is that of the role of IS in attentional tasks. Can we better explain why deliberate IS helps us in some cognitive tasks? And do mind wandering or excessive IS impair task performance? When and how does the switch from cognitive task to IS occur? Can we train to refrain from engaging in excessive IS, can rumination in depression or AVH in schizophrenia be reduced?

The INNERSPEECH project has the ambition to provide a thorough theoretical account of IS, by addressing these 4 basic questions. In the long term, this project could offer insights into the mechanisms inducing mental rumination and AVH, two disorders of IS.

To this aim, we will implement a multi-paradigm approach, gathering data from functional magnetic resonance imaging (fMRI), intracranial electroencephalography (iEEG), transcranial magnetic stimulation (TMS) coupled with electromyography (EMG) and EEG, and behaviour. Five tasks will be undertaken over 4 years. Task 1 is devoted to project management. Task 2 consists in optimizing the multi-paradigm design. In Task 3, fMRI studies will be performed during various types of IS, problem solving and resting states. Behavioural, functional and connectivity analyses will address our 4 basic questions. Task 4 examines the neural dynamics of IS using iEEG. It will provide insight as to where activation stems from (motor vs sensory cortex) during IS. Analyses of frequency of oscillation and region synchronisation will reveal brain dynamics and how transitions between cognitive tasks and spontaneous IS occur. Task 4 will address What, How and What purpose. Task 5 investigates causality. TMS excitation and inhibition, with EEG-EMG recording, will be run to test which of motor and/or sensory activities are causal in IS production and agency monitoring. Task 5 will address What and Who.

Our consortium is optimally composed to address these ambitious issues. It brings together specialists of fMRI (LPNC, UMS IRMaGe, GIPSA-lab, Grenoble), connectivity (CNoS, U. British Columbia, Canada), iEEG (INSERM U1028 Brain dynamics and cognition, Lyon; Labo. Neurophysiopathologie de l'épilepsie, Grenoble U. Hospital), TMS and Human Electrophysiology (Neurolab, Ferrara U., Italy; LPNC), source/agency monitoring (CNoS) and IS characterisation (LPNC, GIPSA-lab, U1028). The project will benefit from the geographical proximity between partners in Grenoble, Lyon and Ferrara and from previous successful collaborations between partners in Grenoble and Lyon as well as Grenoble and Vancouver.