A pilot study into a role for subvocal speech production in the perception and reduction of AVH

Around 16% of the general population hears voices in the absence of speaker.
This is called an Auditory Verbal Hallucination (AVH). For part of the
hallucinating population the hallucinations cause great distress, lead to a
lower global functioning, and result in psychiatric disorder. Despite the high
prevalence of AVH, the underlying cognitive mechanisms are largely unknown. A
promising cognitive theory regarding the origin of AVHs is the *deficit in
forward model production* theory. Deficits in the production of the forward
models results in self-produced actions being perceived as if some external
force controls them. In the case of auditory hallucinations, it is hypothesized
that self-produced speech is perceived as speech produced by someone else due
to the lack of attenuation.One avenue of support for this theory comes from
research in the 1950*s demonstrating that during the perception of AVH,
involvement of articulatory muscles was observed through EMG. The speech
remained inaudible unless amplified (subvocal speech). Importantly, although
the speech production was measurable, the person perceiving the AVH did not
perceive the speech as self-produced. The current proposal seeks to decrease
AVH by training the participant to interrupt the production of subvocal speech.
In a first step we will test whether subvocal speech production is measurable
through EMG, and different from people who do not experience AVH. In a second
study we will visualise the subvocal speech by measuring the activation in
muscles engaged during AVH with EMG, and displaying that muscle activation in
real time on a screen. The participant is trained to link the visual display
with their own muscle activity during humming, whistling and speaking. In a
next step the participant sees the engagement of the muscles during AVH, and is
trained to interrupt the AVH by performing the tasks that use the same muscles
or by intentionally relaxing these muscles. Through the visual feedback the
participant can again identify a role for their own articulators during AVH,
bypassing the failure in the forward model production system, and regain
control. Having a sense of control over AVH is a major improvement in the lives
of these participants. Relatively short training on articulatory muscles with
similar biofeedback mechanisms has been shown to have lasting effects.

The current research project investigates whether subvocal speech production
can be measured using EMGs during AVH. If such subvocal speech is visible in
the EMGs, as previous research suggests, this visualization can be used to
develop biofeedback training. With biofeedback training the participant sees
the muscle activation, thereby correcting incorrect interpretation. The
participant is trained to use this visualization to stop subvocal speech
production, which in turn results in a halt of the AVH. If the biofeedback
training works as proposed, this technique has the potential to form a novel
therapeutic tool.
The objective of the current research project is to pilot whether EMG-based
biofeedback can be used as a novel therapeutic tool to reduce AVH.

The current study is a pilot study with a two-part design. Part 1 is a case
control observational study, Part 2 is a clinical trial study with a
behavioural intervention

In part 1 we measure EMG activation in articulatory muscles while the
participants passively watch a nature movie without sound. Wet-gel electrodes
are strategically placed on articulatory muscles (orbicularis oris superior,
orbicularis oris inferior and sternocleidomastoid with a bipolar electrode
montage). Additionally two electrodes are placed on a reference site
(brachioradialis) to ensure measured orofacial muscle activations are
associated with speech and not generalized muscle contractions. We also record
EMG activation of articulatory muscles during humming, whistling and reading
out loud. Participants who hallucinate are asked to report the perception of
AVH via a button press.
Before the experimental tasks, a few additional measures of importance to the
interpretation and validity of the findings are taken. Firstly, the
Questionnaire for Psychotic Experiences (QPE) will be administered to assess
the presence, severity, and phenomenology of delusions and hallucinations.
Second, a diagnostic interview (M.I.N.I) will be performed with the
participants to assess the current state of the participant*s mental health,
unless a diagnostic interview was been performed in the month preceding the
testing date, in which case those data will be used. Third, to have an estimate
of the IQ of the participant, they perform the Dutch Adult Reading Test (DART).
Additionally, a shadowing procedure is followed where the participant is asked
to repeat out loud what the voices are telling them, to allow for a
characterisation of the verbal hallucinations.

In part 2 we ask the AVH+ patients from part 1 to return. The setting of the
experiment is identical to session 1. In this session we display muscle
activation of that participant in real time on a screen. The participant is
then trained to link the visual display of EMG activation with his/her own
muscle activity during humming, whistling or speaking. In a next step the
participant and experimenter identify the engagement of the muscles during AVH,
and the participant is trained to interrupt the AVH by performing the tasks
that use the same muscles or by intentionally relaxing the muscles. The outcome
measure of this task is the number of successful AVH interruptions as function
of the number of interruption attempts.

The current study aims to investigate whether biofeedback can be used to
interrupt AVH perception. Part 1 of the study observes which, if any, sites
demonstrate a correlation with EMG activity during AVH perception. In part 2 a
behavioural intervention takes place in the form of biofeedback training.
During the session participants are seated in a comfortable chair. Electrodes
are placed on their articulatory muscles. A real-time visualization of the EMG
activation is displayed on a monitor in front of the participant (biofeedback).
The experimenter explains the visualization to the participant. The participant
performs simple behavioural tasks (relaxing, humming, whistling, talking out
loud) to see how this affects the visual display. The biofeedback is
subsequently used to interrupt AVH. The experimenter asks the participant to
indicate the experience of AVH. The experimenter demonstrates to the
participant how the AVH is linked to the visualisation. Experimenter and
participant discuss strategies to engage the AVH relevant musculature using a
behavioural task, and put this in practice. If successful, the participant
continues the training using this strategy. If unsuccessful after multiple
attempts a different strategy is employed; either the behavioural engagement
task is changed or instead of engagement, the relaxation of relevant
musculature is trained.
The experimenter notes which strategies were successful, how many attempts to
interrupt AVH were made, and how many were successful.

The current study sets out to test whether a novel therapeutic tool can be
developed that would give patients who experience AVH more control over the
perception. To this aim, it is necessary that we test participants who
experience AVH. AVHs are most commonly associated with schizophrenia, but are
also observed in other psychiatric disorders including bipolar disorder,
post-traumatic stress disorder, in some neurological disorders (e.g., temporal
lobe epilepsy) and in healthy individuals. For the current study to be
successful it is necessary to recruit participants who perceive AVH at a high
frequency. As a result, a substantial of participants in the current study will
be psychiatric patients.

The number of participant visits are limited (2 max.) and mainly requires time
investment for questionnaires and behavioural testing sessions.There are no
risks associated with taking part in this study beyond normal every-day living.

Potential benefits for the participants who experience AVH is that if the
mechanism and therapy work as proposed, the participant might be able to have
more control over the AVH perception, however this is uncertain. Participants
who do not experience AVH will have no personal benefit from participation.