Swallow-related Brain Activity in Post-total Laryngectomy Patients: A Case          Series Study

Akari Ogawa; Satoko Koganemaru; Toshimitsu Takahashi; Yuu Takemura; Hiroshi Irisawa; Kazutaka Goto; Masao Matsuhashi; Tatsuya Mima; Takashi Mizushima; Kenji Kansaku

doi:10.2490/prm.20230026

ABSTRACT

Background: Total laryngectomy is a surgical procedure to completely remove the hyoid bone, larynx, and associated muscles as a curative treatment for laryngeal cancer. This leads to insufficient swallowing function with compensative movements of the residual tongue to propel the food bolus to the pharynx and esophagus. However, the neurophysiological mechanisms of compensative swallowing after total laryngectomy remain unclear. Recently, swallowing-related cortical activation such as event-related desynchronization (ERD) during swallowing has been reported in healthy participants and neurological patients with dysphagia. Abnormal ERD elucidates the pathophysiological cortical activities that are related to swallowing. No report has investigated ERD in post-total laryngectomy patients.

Case: We investigated ERD during volitional swallowing using electroencephalography in three male patients after total laryngectomy for laryngeal cancer (age and time after surgery: Case 1, 75 years, 10 years; Case 2, 85 years, 19 years; Case 3, 73 years, 19 years). In video fluorographic swallowing studies, we observed compensatory tongue movements such as posterior–inferior retraction of the tongue and contact on the posterior pharyngeal wall in all three cases. Significant ERD was localized in the bilateral medial sensorimotor areas and the left lateral parietal area in Case 1, in the bilateral frontal and left temporal areas in Case 2, and in the left prefrontal and premotor areas in Case 3.

Discussion: These results suggest that cortical activities related to swallowing might reflect cortical reorganization for modified swallowing movements of residual tongue muscles to compensate for reduced swallowing pressure in patients after total laryngectomy.

INTRODUCTION

Total laryngectomy is a surgical procedure for laryngeal cancer that completely removes the hyoid bone, larynx, and associated muscles.¹⁾ This causes anatomical and physiological changes in pharyngeal structure and motion that affect swallowing function. Generally, in the process of swallowing in healthy subjects, the bolus of food is propelled to the esophagus by swallowing pressure produced by tongue movements, laryngeal elevation, and contraction of the cricopharyngeal musculature in the pharyngeal phase.²^,³⁾ However, in post-total laryngectomy patients, the propulsion force for the bolus decreases because of the absence of adequate pressure with laryngeal elevation.⁴^,⁵⁾ Therefore, it is considered that the bolus is driven to the esophagus by the ejection action of the tongue root and by gravity in post-total laryngectomy patients.³⁾ These factors cause prolonged pharyngeal transit duration and residual bolus in the pharynx, resulting in swallowing dysfunction in post-total laryngectomy patients with laryngeal cancer.⁶⁾ Furthermore, most laryngeal cancer patients receiving total laryngectomy in Japan are older than 65 years of age (Cancer Statistics in Japan, https://ganjoho.jp/reg_stat/statistics/stat/cancer/11_larynx.html). Age-related alterations in swallowing function include prolongation of pharyngeal transit duration, delayed laryngeal elevation, delayed opening of the upper esophageal sphincter because of downward deviation of the larynx, and decreased elasticity of the cricopharyngeal musculature.⁷^,⁸⁾ These age-related physiological changes increase the risk of aspiration. However, less attention has been paid to postoperative swallowing dysfunction because there is no risk of laryngotracheal aspiration because of the complete separation of the respiratory and digestive passages. Therefore, the physiological mechanisms of swallowing in post-total laryngectomy patients are still not fully understood. However, it has been reported that more than half of postoperative patients with laryngeal cancer perceive dysphagia and decreased quality of life (QOL).⁹^,¹⁰^,¹¹⁾ Post-total laryngectomy patients in particular experience decreased QOL because of limitations in food morphology and because of the importance of eating in public spaces for social participation.¹²^,¹³⁾ Therefore, it is important to elucidate the physiological mechanisms of swallowing in post-total laryngectomy patients to develop new rehabilitation therapies that can improve their QOL.

In previous studies using positron emission tomography and functional magnetic resonance imaging (fMRI), cortical activation was observed in the primary motor (M1), somatosensory, supplementary, and premotor cortices during volitional swallowing in healthy subjects.¹⁴^,¹⁵⁾ With age-related alteration, the cortical activation, especially in the somatosensory cortex, becomes more symmetrical and extensive during volitional swallowing.¹⁶^,¹⁷⁾ We have also successfully recorded event-related desynchronization (ERD) using electroencephalography (EEG) in the beta frequency band (15–25 Hz) in the bilateral sensorimotor, premotor, and inferior prefrontal areas in healthy people, and the abnormal distribution in patients with amyotrophic lateral sclerosis (ALS).¹⁸^,¹⁹⁾ ERD is a power decrease in the alpha and beta frequency bands during voluntary movements when compared with the power during rest (non-movement)²⁰^,²¹^,²²^,²³^,²⁴⁾ and reflects cortical activation during voluntary movements.²²⁾ In previous studies, ERD was observed in the alpha and beta frequency bands during volitional swallows.²⁵^,²⁶^,²⁷^,²⁸^,²⁹⁾ The emergence of ERD was changed in ALS patients with dysphagia,¹⁹⁾ patients with Parkinson’s disease without dysphagia,²⁶⁾ in post-stroke patients with dysphagia after transcranial direct-current stimulation therapy,²⁹⁾ and in healthy subjects after electrical stimulation of the pharynx.²⁸⁾ These findings suggest that ERD within the beta band frequency is an index of swallowing-related cortical activities.

However, no report has investigated cortical activity during volitional swallowing in post-total laryngectomy patients. We hypothesized that ERD within the beta band frequency during swallowing could be found in cortical areas different from the medial sensorimotor areas in post-total laryngectomy patients. Such activity would suggest brain reorganization to compensate swallowing function by tongue movements after removal of the larynx. Detection of cortical activity related to volitional swallowing would confirm not only swallowing-related cortical function in post-laryngectomy patients, but also the neurophysiological mechanism of compensatory swallowing after removal of the larynx. In this case report, we used EEG to investigate cortical activity related to volitional swallowing in three post-total laryngectomy patients with laryngeal cancer.

CASES

Case Description

Case 1 was a 75-year-old man. At the age of 65 years, he felt hoarseness without swallowing difficulty and 4 months later, he was diagnosed with laryngeal cancer (cT3N0M0). Total laryngectomy and left radical neck dissection (II–V) were undertaken for laryngeal cancer (pT4N0M0) 4 months after the diagnosis. Chemotherapy with tegafur, gimeracil, and oteracil potassium was applied for 2 weeks before surgery, but there was no postoperative radiation or chemotherapy treatment. The assessment was performed 8 years and 11 months after the surgery. Case 2 was an 85-year-old man. Total laryngectomy and radical neck dissection were undertaken at the age of 66 years for laryngeal cancer (T3N1M0). He had no history of chemoradiotherapy for laryngeal cancer. The assessment was performed 19 years and 2 months after the surgery. Case 3 was a 73-year-old man. At the age of 54 years, he underwent total laryngectomy and left radical neck dissection for laryngeal squamous cell carcinoma (T3N1M0). No chemoradiotherapy was given before or after surgery. The assessment was performed 18 years and 11 months after the surgery (Table 1). All three patients had normal cognitive function and had no medical treatment, pre-existing condition, or comorbidity that would impair swallowing function. They showed neither sarcopenic nor aging-related dysphagia before and after surgery. The patients ate a regular diet and cut food into bite-size pieces when necessary. They were able to drink normal water by bending their neck backward without any need to thicken the water. They were all independent in their activities of daily living. This study was approved by the Committee of Medical Ethics of Dokkyo Medical University, Mibu, Japan (No. 30008). The patients provided written informed consent to participate in the study.

Table 1. Clinical findings of cases

Case	Timeline (years)	Clinical findings
1	0 (onset)	Diagnosed with laryngeal cancer (cT3N0M0)
	0.3	Chemotherapy with the compound drug of tegafur, gimeracil, and oteracil potassium for 2 weeks
	0.4	Underwent total laryngectomy and left radical neck dissection (II–V)
	8.0	Evaluated VFSS and EEG
2	0 (onset)	Diagnosed with laryngeal cancer (T3N1M0), underwent total laryngectomy and radical neck dissection
2	19.2	Evaluated VFSS and EEG
3	0 (onset)	Diagnosed with laryngeal squamous cell carcinoma (T3N1M0), underwent total laryngectomy and left radical neck dissection
3	18.1	Evaluated VFSS and EEG

In all cases, total laryngectomy was performed through a U-shaped incision. Post-laryngectomy, the pharynx was reconstructed with mucosal sutures and a permanent tracheal stoma was formed. Postoperative fistula formation was not observed. None of the patients underwent postoperative swallowing rehabilitation. Before EEG, a video fluorographic swallowing study (VFSS) was performed in a relaxed sitting position to confirm tongue movements during swallowing after removal of the larynx. The patients volitionally swallowed 3 mL of water that was injected by the experimenter using a syringe. In all cases, the tongue was elevated in the oral phase and the dorsum of the tongue was pulled inferoposteriorly to contact and press against the posterior wall of the pharynx to generate positive pressure, propelling the bolus into the upper esophagus in the pharyngeal phase as previously reported as compensatory behavior (Fig. 1 showing Case 2).³^,⁴^,⁵⁾ No pharyngeal residue was observed in each case. In VFSS, the swallow reaction time from the arrival of the bolus head at the hypopharynx to the backward retraction of the base of the tongue was 310, 1200, and 260 ms in Cases 1, 2, and 3, respectively. The time from the tail of the bolus passing through the upper esophagus to the return of the tongue base was 310, 410, and 2200 ms in Cases 1, 2, and 3, respectively.

Fig. 1.

VFSS in Case 2 showing the oral preparatory (left) and pharyngeal (right) phases. Black arrow indicates dorsum of the tongue; white arrow indicates posterior wall of the pharynx. In the pharyngeal phase, the dorsum of the tongue was pulled downward and backward to press against the posterior wall of pharynx and to close the hypopharyngeal space.

EEG and Electromyography

The patients were comfortably seated with the upper body raised to around 90° in an armchair during recording. We recorded EEG signals with 32 electrodes during volitional swallowing. EEG signals were recorded using eego^TM sports active electrodes (ANT Neuro, Hengelo, Netherlands) attached inside the EEG cap according to the 10–20 international electrodes system. The CPz location was used for the reference electrode. The EEG signals were amplified with an eego^TM sports amplifier (ANT Neuro). The impedance of all electrodes was less than 15 kΩ³⁰⁾, and the data were sampled at 1 kHz. While recording EEG signals, the experimenter injected 3 mL of water into the patient’s oral cavity via a 3.3-mm i.d. flexible plastic tube every 20 s. If necessary, the patient was allowed to take a break before the next injection. The tip of the tube was placed between the buccal side of the teeth and the cheek, and the tube was fixed on the skin using tape according to methods used in previous magnetoencephalography (MEG) studies.²⁷^,²⁸^,²⁹⁾ We instructed patients to perform volitional swallowing without any head movement until at least 3 s after the water was injected.¹⁸⁾ The patients were also instructed not to swallow saliva after water injection until the initiation of volitional swallowing. We confirmed compliance by visually monitoring the oro-pharyngeal movements and the EMG recording. Volitional swallowing was repeated for 1 h. After swallowing, each patient was permitted as take a break if they wished to do so. The frequency of swallowing depended on the length of the rest periods.

For electromyography (EMG), we used two pairs of bipolar silver/silver chloride electrodes connected to a bipolar eego^TM sports amplifier (ANT Neuro). The electrodes were attached on the anterior point one-third of the way between the mandibular angle and the ridge of the muzzle of both sides where we confirmed muscle contraction on the skin. We concurrently recorded the swallowing movements using a triple-axis accelerometer attached to the anterior part of the patient’s neck based on previous studies.²⁷^,²⁸^,²⁹^,³¹⁾ During volitional swallowing, head movements were also observed with two cameras.

Data Analysis

Preprocessing of EEG and EMG Signals

We performed independent component analysis (ICA) using the EEGlab MATLAB toolbox (MathWorks, Natick, MA) to remove blinking artifacts and electrooculographic activity from the EEG signals. We identified the onset and the completion of each swallowing task from the EMG signals of the intrinsic tongue muscles to segment the EEG signals, similar to previous MEG studies.²⁷^,²⁸^,²⁹^,³²⁾ The onset on the EMG signal (M1) was defined as the time at which the amplitude or frequency increased by more than 100%. The completion on the EMG signal (M2) was defined as the time when the amplitude or frequency of the averaged EMG signal decreased by more than 50%, in accordance with previous studies. Moreover, we manually defined M0 as the time when visible EMG activity at the beginning of swallowing was confirmed to measure the duration of swallowing (from M0 to M2). We defined the onset of swallowing (E1) as M1−0.4 s and the completion of swallowing as M2 to analyze the event-related EEG signals (Fig. 2), as in previous studies.²⁸^,³²^,³³⁾ The time intervals were determined as follows:

Fig. 2.

Definition of onset and completion of volitional swallowing according to swallow-related muscle activities in Case 2. M0 is defined as the time of the first electromyographic activity upon initiating swallowing, M1 is the time at which the main muscles are activated, and M2 is the time to return to baseline.

(1) Execution stage: −0.4 to 0.6 s in reference to M1 (from E1 to E1+1 s)

(2) Resting stage: 0–1 s in reference to M2

EEG Analysis

Using fast Fourier transform (FFT), we calculated the power spectral density of the EEG denoised by ICA for the execution and resting stages. In the execution stage, FFT was applied to 1000-ms segments spanning from the onset of swallowing (from E1 to E1+1.0 s). The baseline for ERD analysis was 1000 ms at rest (from M2 to M2+1.0 s). The upper and lower limits of the FFT were 500 Hz and 1 kHz, respectively. We calculated the average of the power spectral density in the frequency from 15 Hz to 25 Hz in both stages. The ERD was calculated by taking the logarithm of the averaged power spectral density and subtracting that in the resting stage from that in the execution stage. The mean ERD of all channels was obtained for each patient. The 95% confidence limit for EEG power was calculated for the number of trials (n) in each patient as follows³⁴⁾: Confidence limit (95%) = ± (0.851) ∙ (n)^−1/2

To investigate the frequency and temporal properties during swallowing, we performed time–frequency analysis using short-time FT. The FT size was 200 points and the time shift was 50 ms. The analyzed time period was from 1 s before the start of the execution stage (E1−1 s) to 2 s after the start of the execution stage (E1+2 s) and was averaged across all the swallow trials. The data of the channels showing significant ERDs during swallowing were averaged in each case.

RESULTS

No adverse events occurred during the EEG recordings. The patients did not execute double or multiple swallowing and did not swallow saliva during volitional swallowing tasks. They did not get tired of swallowing during the EEG recordings.

During EEG recordings, the number of volitional swallows was 127 in Case 1, 85 in Case 2, and 116 in Case 3. The mean swallow duration was 3.1 ± 1.6 s in Case 1, 2.6 ± 1.2 s in Case 2, and 3.1 ± 1.4 s in Case 3. Inspection of video recordings of the trials confirmed no head movements during EEG recording or during volitional swallowing. Figure 3a shows the topographic mapping of the ERD or event-related synchronization (ERS) averaged over 1 s in the activation phase. The 95% confidence limit for the ERD was −0.0755 for Case 1, −0.0923 for Case 2, and −0.0790 for Case 3. Therefore, ERD smaller than these limits was judged to be non-zero. In Case 1, significant ERD was found in the bilateral medial sensorimotor areas (corresponding to FC1, Cz, and CP2) and the left lateral parietal area (corresponding to CP5). In Case 2, significant ERD was found in the middle and bilateral inferior frontal areas (corresponding to FC5, F7, Fz, and F8) and the left temporal area (corresponding to T7). In Case 3, significant ERD was found in the left prefrontal area (corresponding to Fp1), the right superior premotor area (corresponding to F4), and the bilateral somatosensory areas (corresponding to CP1 and CP2). Figure 3b shows the temporal power modulation during the time period from 1 s before the start of the execution stage to 2 s after the start of the execution stage. The data were averaged in the channels showing significant ERDs in each case (Case1: FC1, Cz, CP2, and CP5; Case 2: FC5, F7, Fz, F8, and T7; Case 3: Fp1, F4, CP1, and CP2). The ERDs emerged immediately before the activation stage and were maintained during the execution stage within the beta frequency range in all the cases.

Fig. 3.

Topographic mapping and time–frequency representation of the mean ERD/ERS in all patients. (a) Topographic mapping display of the swallow-related ERD and ERS. Channels with significant ERD are indicated by yellow circles. Significant ERD is shown in the medial sensorimotor and left lateral parietal areas in Case 1, in the bilateral medial and inferior frontal areas in Case 2, and in the left prefrontal, right premotor, and bilateral somatosensory areas in Case 3. (b) The temporal modulation of the averaged ERD and ERS in each case.

DISCUSSION

In this case study, we investigated the volitional swallowing-related cortical activation in three post-total laryngectomy patients using EEG. Significant ERDs were observed in multiple cortical areas during volitional swallowing but in spatial distributions that were different from the ERDs of healthy subjects. All patients showed cortical activation in both frontal areas involved in a cognitive effort during swallowing¹⁹⁾ and sensorimotor areas that were active in voluntary swallowing in healthy people,¹⁸⁾ although the distribution was different for each case. These patterns of cortical activation after total laryngectomy might suggest the existence of divergent and multiple paths of neural recovery processes among individuals with swallowing dysfunction.

In previous studies, the medial sensorimotor areas were reported to prominently show ERD during volitional swallowing in healthy subjects.¹⁸^,²⁵^,³²⁾ In this study, significant ERD in the medial sensorimotor areas was observed only in Case 1.

The left lateral parietal area in Case 1 also showed ERD. A previous study showed remarkable activation in the left lateral parietal area including angular and supramarginal gyri during effortful swallowing to squeeze the muscles of the back of the tongue rather than usual volitional swallowing.³⁵⁾ Therefore, the ERD in the left lateral parietal area in Case 1 suggests that the patient may have squeezed the muscles of the back of the tongue as performed in effortful swallowing. Although he reported no effort to swallow, it is possible that effortful swallowing-like neural activation occurred because the decrease in swallowing pressure after removal of the pharyngeal muscle elicits an increase in contraction of the muscles of the dorsal surface of the tongue.³⁾

In Case 2, we found significant ERD in the bilateral middle and inferior frontal areas. In previous fMRI studies, the frontal operculum in the inferior frontal areas and the superior premotor cortex in the middle frontal area were reported to show specific activation for swallowing, not for tongue movement, and to play a part in swallowing networks in healthy subjects.³⁶^,³⁷^,³⁸⁾ Therefore, the current finding may suggest enhanced swallowing function, possibly supporting the residual tongue muscles: the genioglossus muscle from the mandible to the tongue and musculus styloglossus, which are able to pull the tongue backward and downward. The left temporal area also showed significant ERD, which might reflect swallowing difficulty after the removal of the larynx because previous reports showed that the temporal region was related to swallowing function and its activity was increased in patients with dysphagia,³⁹^,⁴⁰⁾ even though the patient claimed he was not aware of any cognitive effort during swallowing. Furthermore, the patient in Case 2 showed a delayed swallow reaction time in VFSS. This result may have been related to the frontal and temporal activities with weakened sensorimotor activity, suggesting that additional force may have been required to support the residual tongue muscles in Case 2.

Significant ERD in the left prefrontal area was found in Case 3. In previous studies, ERD in the prefrontal area during volitional movement tasks reflected the cognitive function related to movement tasks such as attention and estimation of the time to control the movements.⁴¹^,⁴²^,⁴³⁾ It has also been reported that activation in the prefrontal area increased depending on the effort during swallowing.³⁵⁾ We have also found ERD in the prefrontal area during volitional swallowing in ALS patients with dysphagia with decreased swallowing pressure.¹⁹⁾ Given that the patient in Case 3 also reported that he made no effort in swallowing, ERD in the prefrontal area may reflect unconscious cognitive process⁴⁴^,⁴⁵⁾ to achieve more forceful tongue movements with decreased propulsion after removal of the pharyngeal muscles. The superior premotor area and the bilateral medial somatosensory areas also showed significant ERD. Brain activation in these areas was shown during voluntary swallowing in healthy people in previous studies.³⁷^,⁴⁶^,⁴⁷⁾ This might reflect swallowing function recovered by residual tongue muscles. In VFSS, the patient in Case 3 showed delayed return of the tongue base after the bolus passed through the upper esophagus. This may have been related to the additional prefrontal activity with sensorimotor activity, suggesting that overactivity of the prefrontal area may delay the relaxation of the tongue in Case 3. In summary, these divergent patterns of cortical activation after total laryngectomy might suggest the existence of divergent and multiple paths of neural recovery processes among individuals with swallowing dysfunction. In terms of cortical reorganization, it is notable that we found significant ERD in the bilateral medial sensorimotor areas only in Case 1 but not in Cases 2 or 3. After total laryngectomy, most pharyngeal muscles are removed along with the hyoid bone¹⁾ and the residual tongue muscles play a part in propulsion of the bolus to the upper esophagus in the pharyngeal phase, like normal pharyngeal muscles.³^,⁴⁸⁾ The somatotopy of the pharyngeal muscles is arranged in a more medial part than that of the tongue muscles according to previous MEG and transcranial magnetic stimulation studies.²⁵^,⁴⁹⁾ If the lack of ERD over these areas is the consequences of the post-surgical loss of laryngeal muscles, the finding in Case 1 may indicate that the medial sensorimotor areas having previously controlled the pharyngeal muscles may control tongue muscles that are compensating for the removed pharyngeal muscles. This may suggest a means of cortical reorganization for functional recovery.

Although the patients in all three cases showed cortical activation in both frontal areas similar to that during effortful swallowing, it is difficult to distinguish which act required more effort: the task itself or the patient’s ability to evoke swallowing. Comparison with age-matched healthy older people in the same regimen would be necessary to clarify the influence of the task.

Physiological swallowing function is also affected by aging. Age-related alterations in swallowing function include prolongation of pharyngeal transit duration, delays of the laryngeal elevation and opening of the upper esophageal sphincter because of downward deviation of the larynx, and decreased elasticity of the cricopharyngeal musculature.⁷^,⁸⁾ To compensate for these alterations, the cortical activation in the somatosensory cortex is expanded.¹⁶^,¹⁷⁾ In addition, these cortical activations are more symmetrically distributed in older subjects than in healthy young subjects.¹⁷⁾ Although the patients in the present study were also older adults, the ERD distribution differs from the cortical activation patterns of older healthy subjects, suggesting that this reflects the effects of modified swallowing movements and cortical reorganization caused by total laryngectomy rather than age-related alteration.

In this study, we characterized the distribution of brain activity during volitional swallowing in three patients with post-total laryngectomy. Although VFSS showed kinematically similar swallowing by tongue movements, the cortical distributions of brain activity during swallowing were different across the cases. This suggests that different processes occur in individual patients. Further studies of a large number of post-total laryngectomy patients and comparison with normal subjects are needed to investigate the mechanism of swallowing and to develop therapeutic interventions for dysphagia after total laryngectomy.

ACKNOWLEDGMENTS

The authors thank A. Sugawara and M. Hayakawa for their technical assistance in the experiments. The authors also thank K. Shimizu for her assistance in caring for the patients. This work was supported by a Grant-in-Aid for Exploratory Research (20K21770), Grants-in-Aid for Scientific Research (B) (21H03308) (SK), Grants-in-Aid for Scientific Research (A) (19H01091, 23H00459) (T. Mima), and Grants-in-Aid for Scientific Research (A) (19H01126) and (B) (19H03939) (K. Kansaku) from the Japan Society for the Promotion of Science (JSPS).

CONFLICTS OF INTEREST

Satoko Koganemaru is affiliated with an endowed department supported by the Kodama Foundation. Masao Matsuhashi is affiliated with the Department of Epilepsy, Movement Disorders and Physiology, Kyoto University, which is part of the Industry–Academia Collaboration supported by Eisai Co. Ltd., Nihon Kohden Corporation, Otsuka Pharmaceutical Co. Ltd., and UCB Japan Co. Ltd. Tatsuya Mima and Kenji Kansaku have research contract funds allocated for medical and science research (≥1M JPY) from Sumitomo Pharma Co. Ltd. (T. Mima) and JSPS, KAKENHI (K. Kansaku).

REFERENCES

Corresponding author

Register with J-STAGE for free!