Introduction
In children and adolescents, traumatic brain injury (TBI) is the leading cause of acquired disability and death [
1]. Importantly, due to the developing brain of a child, brain injuries may have a more devastating impact on children than injuries of the same severity have on adults, thus highlighting the need for additional and supplementary treatment options to improve prognosis especially in this population [
2,
3]. Additionally, long reconvalescence after severe TBI and need for rehabilitation programs are associated with high healthcare costs and, thus, improving treatment after TBI is of high societal relevance [
4].
Many TBI patients show electroencephalographic (EEG) slowing characterized by predominant delta waves (oscillations with a frequency of around 0.5–4 Hz; 5,6). These brain patterns are considered markers for neuroplasticity [
7,
8]. Some literature describes a potential link with clearance of the brain from metabolic waste products and synaptic plasticity especially important after brain injury [
9‐
13]. Therefore, boosting delta waves might represent a promising approach. Indeed, in animal studies (rats), interventions that enhance delta waves after TBI improved neurological and behavioural outcome [
14]. Likewise, in clinical patient studies, delta waves were positively linked with outcome measures in patients after brain injury or after cardiac surgery [
15‐
17]. However, EEG slowing has also been associated with prolonged TBI patient recovery in the clinic [
5]. Taken together, evidence is thus far inconclusive: Potentially, delta waves represent neurological recovery [
7] and might accordingly be more pronounced in patients with more severe brain injuries (higher need for neuronal reorganization), therefore indirectly correlating with worse outcome. However, causal evidence is scarce and, thus, experimental modulation of delta waves in the acute stage after severe TBI is needed to investigate their functional relevance. Admittedly, it is a big step from preclinical evidence to application in acute pediatrics, but direct translation potentially allows to prevent unnecessary delays.
Apart from the intention to improve neurodevelopmental outcome of children after severe TBI, a non-pharmacological approach would be very welcome. The analgosedative medication used in daily clinical practice go along with side-effects and withdrawal that should be minimized. Various non-pharmacological methods to boost delta waves are available, the most promising being auditory phase-targeted stimulation [
18]: Brain activity is measured using EEG to detect delta waves. Upon recognition, brief sounds are presented targeted to the up-phase of the delta wave (up-going slope in the positive half wave). Many studies replicated efficacy of this technique to boost delta waves [
18‐
21]. However, auditory stimulation has so far only been investigated during natural sleep. Application in the acute stage after severe TBI in analgosedated, muscle relaxed, and ventilated patients in an intensive care unit has yet to be evaluated. There is no research to this issue, therefore, little knowledge exists. Maybe our feasibility study will contribute to pave a way for an approach to investigate the potential of auditory phase targeted stimulation in this vulnerable patient group.
Therefore, as an important first step in the vision of having non-pharmacological intervention options to improve neuroplasticity, we conducted a feasibility study to investigate technical feasibility and tolerability of applying a device capable of delivering auditory stimulation in TBI patients treated in the pediatric intensive care unit (PICU). After investigating feasibility of continuously recording signals such as EEG needed for auditory stimulation over time in the PICU and adaption of the stimulation algorithm, feasibility of applying tones to enhance delta waves was examined.
Methods
Patients and setting
Pediatric patients (age ≥ 12 months) with severe TBI requiring management in the PICU qualified for inclusion in the study (TBI classification according to Glasgow Coma Scale score < 8; 22,23). Patients were treated according to the international guidelines by Kochanek and colleagues [
24]. All these patients were comatose due to the traumatic brain injury and treated with medication for sedation, analgesia, and muscle relaxation in varying dosage depending on the patient’s needs to be in a clinically stable status without increased intracranial pressure. One legal representative was required to have good German knowledge. Exclusion criteria were known deafness, neurological or syndromal pre-existing condition, bilateral basal skull fracture, serious skin problems in the face/ear area, and fulfilment of clinical criteria of brain death within 24 h after hospital admission.
The feasibility study was conducted on a level III PICU (27 beds) of the University Children’s Hospital Zurich (Switzerland) between May 2019 and August 2021. All regulations with regards to Covid-19 safety measures were fulfilled at any time.
This feasibility study was approved by the institutional research ethics committee of ETH Zurich (EK 2019-N-18). Informed consent by the parents was obtained.
Procedure
BB (senior PICU consultant) checked eligibility of TBI patients and informed the legal representative(s) about the study. Recordings were started as soon as possible after informed consent was obtained and ended according to patients’ situation. All patients were in supine position, head up 30 degrees. After recording phase, parents and nurses/physicians were asked about perceived disadvantages or benefits for the child due to the study (yes vs. no). Due to the challenging environment, the feasibility study was subdivided into two consecutive phases:
1)
Recording feasibility: First, feasibility of continuously recording signals such as EEG needed for auditory stimulation (see below) over multiple days in the PICU was investigated, but auditory stimulation was muted. Study visits took place about every 24 h to exchange electrodes and devices (because of depleting batteries). Collected data was then used to configure the stimulation algorithm.
2)
Stimulation feasibility: When adaptation of the stimulation algorithm was finished, the second phase started and feasibility of applying tones targeted to the up-phase of delta waves was examined (tolerability and efficacy of tones on delta waves). With a cautious approach, auditory stimulation was enabled for 30–60 min. Criteria to enable the stimulation: Patient is in a stable condition using treatment recommendations for management of TBI by Kochanek and colleagues [
24] and shows similar brain activity as collected in the recording feasibility phase to ensure adequate stimulation algorithm configuration. Therefore, the EEG was inspected in more frequent intervals than the 24 h device exchange. Feasibility was assessed as stimulation tolerability (report of nurses, visual inspection of EEG, heart rate, intracranial pressure) and effect on EEG spectrum. Intolerability based on clinical parameters was defined as augmentation of heart rate and intracranial pressure at beginning of the stimulation and regression to baseline after stimulation offset. Auditory stimulation as every kind of stimulation may produce stress for TBI patients with consecutive increasing intracranial pressure and development of a secondary brain damage. It is crucial to prevent TBI patients of a secondary brain damage to ameliorate neurodevelopmental outcome.
Measures and devices
Patient information was taken from patients’ charts: diagnosis, medication, sedation-agitation scale (SAS; 25) score assessed once every work shift. Additionally, for patients in the stimulation tolerability phase, heart rate and intracranial pressure were extracted from a 24 h window around the stimulation phase in one minute intervals. These clinical parameters are transferred continuously to the PDMS (patient data management system).
Electroencephalography was measured using a mobile, single-channel EEG device (MHSL-SleepBand version 2, sampling frequency 250 Hz; 26) with three self-adhesive electrodes (Ambu
® BlueSensor N; Ballerup, Denmark). The recording electrode was placed around the Fpz location according to the standard 10–20 system [
27], but placement was individually adapted according to patients’ external injuries. Reference and ground electrodes were placed behind the ears (mastoids). The standard setup of the MHSL-SleepBand includes a headband (with embedded electrodes) and additional recording of electrooculography and electromyography. However, due to the challenging situation on the PICU, only the EEG was recorded and electrode cables were fixed with tape because the headband could not be used due to head injuries or the intracranial pressure probe.
External movement of the patients caused by medical or study staff was recorded using an actigraph placed on arm wrist or ankle depending on where the cannulas were inserted (GENEActiv®, Activinsights, Cambridgeshire, UK). Because the PICU represents a noisy environment that might generate confounding factors on a potential stimulation effect, we recorded in addition to the EEG several ambient parameters. Sound level (only loudness (dB), no content) was recorded with a sound level meter (UT352, UNI-T, China). Ambient light levels to characterize the artificial light that prevents natural day-night-rhythm was measured with a mobile device placed at the backside of the patients’ bed (GENEActiv®, Activinsights, Cambridgeshire, UK). All devices except the actigraphy to capture external movement were placed in a small container which was hanged at the backside of the patients’ bed. Additionally, an information sheet for nurses and physicians was placed next to the patients’ bed with a short study description, instructions on how to remove devices, and contact information of the study team.
Auditory stimulation
For this phase, in-ear headphones (Sennheiser, model CX 3.00, Wedemark, Germany) were padded with cotton and fixed with tape to the patients’ ear. Auditory stimulation (50 ms 1/f pink noise bursts, volume about 53 dB SPL) was delivered in alternating 16 s ON (stimulation enabled) and 16 s OFF (stimulation disabled) windows. Pink noise is the most common acoustic stimulus in auditory stimulation studies and efficacy has been previously shown [
18]. Real-time detection of delta waves and phase estimation were autonomously performed by the MHSL-SleepBand version 2 (see Ferster and colleagues [
26] for technical details of the algorithms). The target phase range was in the up-going slope in the positive half-wave of the oscillation (0 to 90°).
EEG preprocessing and statistical analysis
For all analyses, EEG data was notch filtered (50 Hz). For spectral analyses, EEG data was further filtered between 0.5 and 38 Hz and Welch’s power spectral density was estimated in 0.25 Hz bins and 20 s epochs using 4 s hanning windows (no overlap). For phase analysis of applied stimulations, EEG data was filtered between 0.5 and 2 Hz and Hilbert transformation was applied. Circular statistics are presented. EEG data of patient 4 showed pronounced cardiogenic artefacts which were removed using a previously described algorithm (EEG signal filtered between 20 and 40 Hz was used instead of a separate ECG signal; 28).
To assess stimulation effect on EEG, spectrums in ON vs. OFF windows were compared. Spectral analysis was performed in 16 s epochs of stimulation windows (no overlap 4s hanning windows) and unpaired t-tests were used for statistical comparison in each frequency bin.
Discussion
Boosting delta waves might support processes such as brain plasticity and clearance of the brain from toxic waste products that have been linked with those brain patterns [
7‐
13]. The aim of this feasibility study was to investigate feasibility of using auditory phase-targeted stimulation to enhance delta waves in pediatric patients after severe TBI treated on the PICU. To do so, continuous recording of brain activity using an EEG device capable of automatic delta wave detection and stimulation is required. Indeed, we demonstrated good feasibility of recording EEG data over multiple days in four pediatric patients (23 to 88 h of EEG data). After having verified good algorithm performance in the first three patients, stimulation was enabled for 50 min in patient 4 and preliminary evidence for good tolerability of auditory stimulation was obtained: Neither were acute clinical stress reactions reported by nurses or visible in heart rate, blood pressure or intracranial pressure, nor were arousals in the EEG. However, no significant increase of delta power was observed when stimulation was enabled compared to when it was disabled and, thus, the stimulation effect to enhance delta waves remains unclear.
Our data shows that continuous recordings allowing for auditory stimulation over several days are feasible as long as the patient is in a stable clinical condition without increased intracranial pressure. Based on our preliminary evidence it is not possible yet to define whether stable condition of the patient should remain an application criterion for auditory stimulation for future applications. All patients remained intubated and sedated throughout recording periods. Thus, whether recordings could be continued after emergence from comatose state, if not followed by hyperactive delirium, should be evaluated in a further study. The study devices did not interfere with medical procedures except for the actigraph which had to be removed or relocated in all but one patient. Therefore, actigraphy not essential for auditory stimulation should rather be omitted. For all other devices, removal was only necessary in one of the four patients who required a CT scan.
While feasibility of applying devices required for auditory stimulation was thus demonstrated and our very preliminary evidence suggests good stimulation tolerability, effect of stimulation remains unclear and requires further investigation. Thus far, at least three potential explanations for inefficacy in patient 4 are available: First, non-responders have been described as well in other studies where the approach was overall successful to boost delta waves [
30]. Therefore, auditory stimulation might be effective in most, but not all TBI patients, in analogy to what has been shown in healthy sleepers. Second, on a PICU contrary to a lab setting, environment and other conditions are not well-controlled and, thus, we think that effects might be difficult to capture statistically within various other influences such as changes in medication doses and varying background noise. For the latter, evoked auditory potential protocols might provide more insights. Third, delta waves in a comatose and deeply sedated patient might simply be less modifiable than delta waves during natural sleep. This would be in line with a recent study which showed that pharmacologically enhanced delta power with sodium oxybate was not modifiable with auditory stimulation in rats [
31]. Taken together, more research is required to evaluate efficacy of auditory stimulation to enhance delta waves in analgosedated TBI patients in an intensive care environment.
Nonetheless, TBI research in the PICU is a challenging intent: A 24/7 study team is required and close collaboration and exchange with PICU staff is crucial for continuous recordings over multiple days. Special attention should be drawn to electrodes and tape which have to be easily replaceable to avoid stress reactions of patients. Live visualisation of the EEG signal would enable real-time assessment of electrode connection, avoiding data loss due to bad signal quality, and, thus, a device allowing for online monitoring would be preferable. In the meantime, this has already been realized with the more integrated successor of the MHSL-SleepBand version 3 [
32]. However, the biggest challenge, is the low patient numbers: In a period of over two years, only 4 out of 16 patients with severe TBI could be included in the study due to several reasons (Fig.
1). Hence, to gain sample sizes that allow for statistical evaluation of stimulation efficacy and to draw more firm conclusions, a multicentre study with long study duration, together with a flexible study team are required. An international multicentre collaboration might be needed. Taken together, investigating efficacy of auditory stimulation to enhance delta waves after severe TBI is very challenging, though of great potential in the vision of improving prognosis after pediatric TBI. Another complementary approach could be to shift auditory stimulation intervention to the post-acute phase after TBI during rehabilitation phase when medication is reduced and patients start showing natural sleep again rather than in the acute phase of the TBI when treatment on the PICU is required and patients are analgosedated.
Conclusion
Good feasibility of performing unsupervised recordings needed for auditory stimulation in patients with severe TBI on the PICU was demonstrated. Additionally, preliminary evidence for good tolerability of auditory stimuli was obtained, but effect of sounds to modify delta waves remains unclear. Future studies should investigate similarities (and differences) between delta waves in the acute phase after severe TBI, potentially biased or even induced by analgosedation and during deep sleep, to gain more knowledge about modifiability and, in turn, potential of using auditory stimulation to improve recovery from severe TBI. For clinical studies to investigate efficacy of auditory stimulation on patient recovery, we would suggest to shift intervention to rehabilitation phase when patient has emerged from (induced) coma and shows physiological sleep.
Acknowledgements
This work was conducted as part of the Hochschulmedizin Zurich Flagship «SleepLoop». Many thanks go to Sven Leach and Dr Georgia Sousouri for help with data collection, Dr Jelena Skorucak for help with cleaning of EEG data, to Dr Laura Tüshaus for organizational help, and to Drs Maria Laura Ferster and Giulia da Poian for technical support.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.