Abstract

Background:

Neurological decompression sickness (NDCS) is a potentially severe complication of diving, but its effects on autonomic cardiovascular regulation remain poorly understood. Heart rate variability (HRV) has emerged as a noninvasive method for assessing autonomic nervous system dynamics in both physiological and pathological conditions.

Objective:

To describe multidomain HRV patterns in divers presenting with NDCS prior to hyperbaric oxygen therapy (HBOT).

Methods:

This exploratory case series included eight divers diagnosed with NDCS at a tertiary hyperbaric medicine center. Continuous electrocardiographic monitoring was performed before HBOT, and HRV analysis was conducted using a standardized offline processing pipeline. Time-domain (SDNN, RMSSD, pNN50), frequency-domain (HF power, LF/HF ratio), and nonlinear metrics (Shannon entropy and two-dimensional Poincaré analysis) were evaluated. HRV parameters were analyzed both as absolute values and as Z-score normalized values relative to published reference populations.

Results:

Conventional HRV indices demonstrated marked interindividual variability across cases. Some patients exhibited elevated global variability and vagally associated indices, whereas others showed reduced or near-reference values. Entropy-based measures also varied across the cohort, with near-reference values in some cases and lower values in others, particularly Cases 7 and 8. These findings suggest heterogeneous alterations in cardiac rhythm dynamics rather than a uniform HRV phenotype.

Conclusions:

NDCS may be associated with measurable but heterogeneous disturbances in autonomic cardiovascular regulation detectable through HRV analysis. Entropy-based measures may provide exploratory information on NN interval distributional irregularity, but these findings should be interpreted cautiously because of the small sample size, retrospective design, spontaneous respiration, and methodological sensitivity of entropy estimates. Larger prospective studies are needed to determine whether HRV can contribute to physiological characterization or monitoring of decompression illness.

Introduction

Heart rate variability (HRV) is a widely used noninvasive marker of autonomic nervous system regulation in both clinical medicine and physiology research. Across multiple medical fields, HRV has been proposed as a useful tool for physiological monitoring, prognostic assessment, and early detection of disease (Addleman et al., 2025). In acute care settings, reductions in HRV are consistently associated with illness severity and mortality, and short-term HRV measurements have been shown to predict outcomes in conditions such as sepsis and critical illness (Karmali et al., 2017; de Castilho et al., 2018; Liu et al., 2021; Keng et al., 2025; Huang et al., 2025). Continuous HRV monitoring may also precede clinical deterioration by several hours to days (Ahmad et al., 2009; Bodénes et al., 2022).

In diving physiology, immersion and increased ambient pressure induce characteristic autonomic responses, typically including bradycardia and increased parasympathetic modulation related to activation of the diving reflex (Schipke et al., 2001; Lafère et al., 2021; Vulić et al., 2024; Noh et al., 2018). However, environmental stressors, task load, and repeated dives can produce more complex patterns of autonomic activation (Flouris and Scott, 2009; Schaller et al., 2021; Berry et al., 2017). Additional studies and reviews have evaluated HRV during underwater exercise, cold-water diving, simulated diving, and recreational diving using continuous ECG or Holter monitoring (Koch et al., 2022; Lundell and Ojanen, 2023; Lundell et al., 2020; Lundell et al., 2021; Rajdeep et al., 2014; Stelmaszczyk et al., 2019).

Compared with physiological diving responses, relatively little is known about autonomic regulation during decompression sickness (DCS). Experimental studies in animal models demonstrate substantial alterations in autonomic nervous system activity during DCS, including impaired autonomic modulation during neurological injury (Bai et al., 2009) and distinct autonomic signatures preceding cardiopulmonary DCS (Bai et al., 2013; Bai, 2011). Human data remain limited and consist primarily of isolated clinical observations and experimental studies of decompression stress (Schirato et al., 2018; Schmitz, 2025).

The objective of this study was therefore to describe heart rate variability patterns in a series of divers presenting with neurological decompression sickness prior to hyperbaric oxygen therapy, using time-domain, frequency-domain, and nonlinear HRV metrics.

Case reports

We report eight adult divers with neurological decompression sickness who met both clinical inclusion criteria and predefined ECG signal-quality criteria for HRV analysis and were evaluated at our hyperbaric facility between the 1st of February and the 9th of December 2025, with full symptomatic resolution after a single hyperbaric oxygen therapy (HBOT) session, as summarized in Table 1. No significant past medical history was reported by any diver and none of the divers reported chronic medication use. Diagnosis of NDCS was established clinically by physicians experienced in diving medicine, based on dive exposure, symptom evolution, neurological findings, and response to hyperbaric oxygen therapy, after exclusion of alternative diagnoses when appropriate. Long-term neurological follow-up was not systematically available. All patients received hyperbaric oxygen therapy following the USN TT6 within 60 minutes of their arrival to the Hyperbaric Service and were evaluated 24 and 48 hours after the treatment to ensure symptomatic resolution.

CaseAgeGenderGas usedSymptomsTime to treatmentImaging126MaleAirLeft paresthesia, vertigo, mild weakness15 hChest X ray normal242MaleAir / surface-supplied airLeft hemiparesis, hypotension, bradycardia72 hAtelectasis353MaleEAN32Right leg monoplegia, cutis, chest pain40 hChest X ray normal429MaleAir / compressor-supplied airSevere Inner Ear Decompression30 hCerebral CT Scan normal555FemaleEAN32Right Lower Extremity Monoplegia112 hNone635MaleEAN32Vestibular Dysfunction and Left Upper Extremity Monoplegia68 hChest X ray normal739FemaleAirLeft hemiparesis with Diplopia25 hNone858FemaleAirRight lower extremity monoplegia with amnesia20 hCerebral CT Scan normal

Summary of eight consecutive analyzable cases of neurological decompression sickness treated with a single session of hyperbaric oxygen therapy.

Case 1 – Cerebral and inner ear symptoms after a single recreational dive

A healthy 26-year-old recreational male diver (10 lifetime dives) completed a 60-min air dive to 20.8 msw without a safety stop. Within 5 min of surfacing, he developed transient cutis marmorata of both upper limbs, resolving spontaneously within 6 h. One hour post-dive, he reported vertigo, lightheadedness, headache, vague left abdominal discomfort, and left-sided paresthesias of the arm, leg, chest, abdomen, and back, with mild left upper-limb weakness.

Fifteen hours later he presented alert (GCS 15) with a right-sided Romberg sign and mild weakness of the left biceps and forearm flexors (4/5). Chest radiograph was normal. Treatment with a U.S. Navy Treatment Table 6 (USN TT6) led to complete resolution of symptoms by the end of the session, and he remained asymptomatic at 24- and 48-h follow-up.

Case 2 – Hemiparesis after surface-supplied commercial diving

A 42-year-old professional surface-supplied male diver, performing 5–10 dives daily for 15 days, on the day of the event conducted a 60-min dive to ~30 msw for net repair. Shortly after surfacing, he developed left hemiparesis, vertigo, right facial palsy, bilateral hip and shoulder pain, hypotension, and bradycardia.

Three days later he presented to a hyperbaric facility with persistent left hemiparesis, musculoskeletal pain, and diffuse 2–4 cm bluish-brown macules on the trunk and upper limbs, compatible with cutaneous DCS. He was alert (GCS 15), hemodynamically stable, with monoparesis of the left upper limb (biceps and triceps 3–4/5) and a right-sided Romberg sign. Brain CT was normal; thoracic CT showed atelectasis without pneumothorax. D-dimer was moderately elevated. After a USN TT6, he was asymptomatic at 24- and 48-h review.

Case 3 – Monoplegia and chest symptoms after repetitive diving

A 53-year-old male divemaster (~380 lifetime dives) completed 3–4 EAN32 dives per day over six consecutive days on a live-aboard, to a maximum depth of 32.7 msw. On day 5 he developed a transient abdominal rash, which recurred on day 6 with diffuse chest and neck pain and mild dysphonia. Four hours of surface oxygen did not relieve symptoms. On return to shore, he noted persistent abdominal cutis marmorata, right-sided chest discomfort, and reduced voice projection.

He presented about 40 h after the last dive with right lower-limb weakness (hip flexors, quadriceps, and extensors 4/5); Romberg sign, ECG, and chest X-ray were unremarkable. A single USN TT6 produced full symptom resolution within 20 min. He remained asymptomatic at 24 and 48 h. Laboratory tests (CBC, D-dimer, troponin, pro-BNP) and contrast transthoracic echocardiography showed no abnormalities or right-to-left shunt.

Case 4 – Severe inner ear decompression sickness in an artisanal compressor diver

A 29-year-old male artisanal compressor diver presented after two days of intensive diving (three dives to 30 m on day 1, followed by two dives to 40 m on day 2, including one of 140 minutes). Approximately 10 minutes after his second dive on the second day he developed left knee pain and bilateral shoulder pain. He attempted in-water recompression at 10 msw, but was forced to perform an emergency ascent due to a hose malfunction; upon surfacing, he developed left-sided vertigo. A second in-water recompression attempt failed due to mask problems. Symptoms progressively worsened, evolving to severe left-sided vertigo with inability to walk, complete left-sided hearing loss, nausea, vomiting, and shoulder pain. On initial evaluation, 30 h after surfacing, he was alert (GCS 15), normotensive, with a positive Romberg (immediate left deviation), comparative left-sided deafness, left horizontal nystagmus, and inability to ambulate, without motor deficits in the extremities.

Case 5 – Right lower extremity monoplegia preceded by recurrent cutaneous manifestations

A 55-year-old female experienced diver (Master Scuba Diver, Nitrox certified, 570 lifetime dives) presented 112 hours after surfacing following a liveaboard trip. Over multiple dive days on EAN32, she experienced recurrent cutaneous DCS (pruritic skin rash on the left breast and abdomen), following an accidental rapid ascent from 12 to 3 msw. Symptoms resolved spontaneously or during subsequent dives on multiple occasions. On her fifth dive day, after three dives (maximum 27 msw), she developed extreme fatigue and phosphenes, which resolved with rest. On the sixth dive day, a single dive to 27 msw triggered recurrence of the abdominal rash, which resolved with 1 hour of surface oxygen; however, back pain subsequently developed and responded to an additional hour of oxygen. On presentation, neurological examination revealed a negative Romberg but significant right lower extremity weakness (hip flexor, gluteus maximus, hamstrings, extensors, and gastrocnemius graded 3/5; quadriceps 4/5), with normal upper extremity strength.

Case 6 – Vestibular dysfunction and left upper extremity monoplegia in a professional dive instructor

A 35-year-old male dive instructor presented 68 hours after surfacing, following seven days of three-dives-per-day on EAN32 (depths ranging 75–104 fsw). On his day 5, after his third dive, he developed a persistent bilateral abdominal skin rash (cutis marmorata) without other symptoms, which did not resolve spontaneously. He continued diving on the last day without incident. On presentation, he additionally reported intermittent right-sided chest discomfort during inhalation underwater, present for several months. Neurological examination revealed a positive Romberg to the left and left upper extremity weakness (deltoids and biceps 4/5, forearm flexor 3/5, forearm extension 4/5), with normal lower extremity strength.

Case 7 – Left hemiparesis with diplopia following repetitive recreational diving

A 39-year-old female recreational diver (AOWD, 40–50 lifetime dives) presented 25 hours after completing three consecutive dives (maximum depths 22–28 m, total bottom time ~158 minutes). She developed epigastralgia, fatigue, diplopia, bilateral elbow pain, and generalized muscle weakness. Neurological examination revealed a positive Romberg (immediate left deviation), right hemiparesis affecting primarily the lower extremity (gluteus maximus, quadriceps, and hamstrings graded 3/5), and mild upper extremity weakness (deltoids and biceps 4/5 on the right).

Case 8 – Right lower extremity monoplegia with amnesia

A 58-year old female recreational diver (Rescue Diver, 78 lifetime dives) presented 20 hours after ten consecutive dives, two per day, all on air and within non decompression limits. About 10 minutes after her last dive she presented skin rash on her abdomen. About 2 hours after the dive she presented an episode of amnesia without loss of consciousness for 20 minutes. Cerebral CT Scan and Chest X-ray were reported as normal. Physical examination showed right lower-extremity monoplegia, abdominal hyperesthesia, psychomotor slowing, impaired abstract thinking, and confusion. All symptoms resolved after a single hyperbaric treatment session.

MethodologyStudy design and setting

This case series describes eight adult divers diagnosed with neurological decompression sickness (NDCS) who were evaluated at a tertiary hyperbaric medicine center in Costa Rica between 1st of February and 9th of December 2025. All patients presented with acute or subacute neurological symptoms following scuba diving and underwent standardized clinical evaluation and neurophysiological monitoring prior to hyperbaric oxygen therapy (HBOT).

Patient identification, routine ECG monitoring, and analytical cohort

This study was designed as a retrospective descriptive case series. The year 2025 was predefined as the review period for identifying eligible patients with neurological decompression sickness evaluated at our hyperbaric medicine center. At our institution, all patients treated for decompression sickness routinely undergo continuous ECG monitoring before, during, and after hyperbaric oxygen therapy as part of the standard clinical protocol to ensure cardiac and hemodynamic stability during treatment. Therefore, the ECG recordings analyzed in this study were obtained as part of routine clinical monitoring rather than as a prospective research intervention.

During the 2025 retrospective review period, nine adult divers with clinically diagnosed neurological decompression sickness underwent pre-HBOT ECG monitoring and were initially considered for HRV analysis. One patient, evaluated chronologically between the current Cases 3 and 4, was excluded from the final HRV analysis because of significant atrial dysrhythmia with excessive irregular RR intervals. This exclusion was applied after inspection of the ECG-derived RR interval series and signal-quality metrics, because the rhythm disturbance precluded reliable normal-to-normal interval reconstruction and could have substantially biased time-domain, frequency-domain, and entropy-based HRV metrics.

The final analytical cohort therefore consisted of eight consecutive analyzable patients who fulfilled both clinical eligibility criteria and ECG signal-quality criteria for HRV analysis. No additional patients from the 2025 review period were excluded because of missing clinical information, unavailable pre-HBOT ECG data, inadequate recording duration, insufficient raw ECG quality, or incomplete HRV processing. Thus, the final cohort represents the consecutive analyzable NDCS cases from the predefined 2025 retrospective review period after exclusion of one patient with clinically significant atrial dysrhythmia.

Clinical severity scoring

Clinical severity was assessed using three previously published decompression sickness scoring systems. The Boussuges score is a neurological prognostic score developed for spinal cord decompression sickness and incorporates the distribution and severity of neurological deficits (range 1–22, higher scores indicating greater severity) (Blatteau et al., 2011). The Perceived Severity Index (PSI) is a probabilistic severity classification system based on symptom burden and estimated DCS severity ranging from 1 to 6 (Howle et al., 2017). The MEDSUBHYP score is a clinical severity scale developed to quantify decompression illness severity across neurological and systemic manifestations (range 0–21, higher scores indicating greater severity) (Blatteau et al., 2011). Scores were assigned based on the initial neurological examination performed prior to HBOT. Clinical severity scores are summarized in Table 2.

CaseBoussuges score (1–22)PSI (1-6)MEDSUBHYP score (0-21)Case 1437Case 2619Case 3628Case 412110Case 512111Case 616113Case 715114Case 81019

Comparison of decompression sickness (DCS) severity across eight clinical cases using three established grading systems.

Decompression stress

Dive exposure was summarized descriptively using the available clinical dive history and, when available, reconstructed dive-profile information. Breathing gas for the relevant index or repetitive dive exposure is reported in Table 1. Estimated decompression exposure was characterized using a custom Bühlmann ZHL-16 implementation, reporting the maximum fraction of the allowable inert-gas gradient reached during the reconstructed profile and the tissue half-time compartment in which this maximum occurred. These decompression variables were used only as descriptive indicators of exposure and were not treated as validated predictors of clinical severity. Estimated decompression exposure variables are summarized in Table 3. Full modeling details are provided in Supplementary Methods S1.

CaseDays of consecutive dives% of maximum gradient factorCompartment of maximum Bühlmann gradient fraction (tissue half-time, min)Case 1152%18.5Case 21579%18.5Case 3672%18.5Case 410139%53Case 5256%37Case 6784%18.5Case 7263%12.5Case 81054%12.5

Dive exposure characteristics and estimated decompression stress using the Bühlmann ZHL-16 model.

ECG monitoring and HRV analysis

Continuous electrocardiographic monitoring was performed for 20 minutes prior to hyperbaric oxygen therapy (HBOT) in a quiet, temperature-controlled room with patients in the supine position. Recordings were obtained during spontaneous breathing without paced respiratory control or respiratory monitoring. Due to the acute clinical setting, factors such as recent food intake, caffeine consumption, circadian variability, pain, psychological stress, and medication exposure could not be fully standardized across patients.

ECG acquisition was performed using a Mindray BeneVision N19 monitor. The sampling configuration provided a temporal resolution of approximately 2 ms, sufficient for accurate R-peak detection and short-term HRV analysis. Raw ECG traces were exported for offline processing using a custom Python-based analytical pipeline (Python version 3.14.3) utilizing the NeuroKit2 and SciPy libraries, following international standards for HRV analysis and interpretation (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, 1996). From each 20-minute pre-HBOT ECG recording, an approximately 10-minute analyzable segment was selected for HRV analysis. The target duration was 10 minutes; however, the final analyzed duration varied slightly across cases because terminal artifacts or non-analyzable edge intervals were excluded during RR/NN reconstruction. The final analyzed duration for each case is reported in Supplementary Table 1. All HRV values reported in Table 4 and all Z-scores shown in Figure 1 were computed from these final approximately 10-minute analyzable segments.

CaseSDNN (ms)RMSSD (ms)PNN50 (%)HF Power (ms2)LF/HFSD1SD2Shannon entropyCase 181.566.343.31056.42.0946.9105.33.09Case 247.135.316.3223.01.3024.961.83.08Case 352.253.241.2510.41.0437.763.62.90Case 428.342.630.5240.60.9730.226.22.89Case 559.662.041.4537.80.5743.872.02.87Case 665.558.840.6907.50.9841.682.72.99Case 712.415.80.427.00.7411.213.42.21Case 845.154.73.9321.10.6338.750.81.59

Z-score normalized heart rate variability (HRV) profiles for eight cases of neurological decompression sickness (NDCS). The bar chart illustrates deviations from normative reference values (Z = 0, blue dashed line). The blue dotted lines mark one standard deviation (Z = ± 1), and the red dotted lines mark two standard deviations (Z = ± 2). Metrics include SDNN, RMSSD, pNN50, HF Power, LF/HF, and Shannon Entropy. Age- and sex-specific reference means and standard deviations used for Z-score computation were obtained from Voss et al. (2015) and are reported in Supplementary Table S2. Entropy-based Z-scores are exploratory because entropy estimates are method-dependent and the reference entropy metric is not identical to the histogram-based Shannon entropy used in the present analysis.

Segment selection was performed manually before HRV metric computation, using the same technical rule in all cases: continuity of the ECG trace, relative signal stability, absence of major movement or signal-loss artifacts, and the lowest visible artifact burden within the available pre-HBOT recording. The segment was not selected on the basis of the resulting HRV values. The analyst was not blinded to the clinical presentation, because this was a retrospective clinical case series and the ECG recordings were reviewed together with the clinical record. All selected segments were obtained after arrival to the hyperbaric service and before HBOT initiation, during routine pre-treatment monitoring; exact segment start times relative to arrival and HBOT initiation were not systematically recorded.

R-peaks were detected from the ECG recordings using a modified Pan–Tompkins-based offline processing pipeline. Raw RR interval tachograms were first generated from the detected R peaks and inspected as part of the signal-quality control procedure. Ectopic or artifact intervals were identified using predefined physiological limits and local deviation criteria. Specifically, RR intervals <300 ms or >2000 ms were classified as nonphysiological. In addition, RR intervals showing >20% relative deviation from the surrounding local median within a moving local window were classified as ectopic/artifact intervals. The resulting classifications were verified by visual inspection of the raw RR tachograms and corrected NN tachograms.

Internal artifact intervals surrounded by valid RR intervals were corrected by interpolation before HRV computation. Terminal artifact intervals located at the beginning or end of the analyzed series, when not reliably bracketed by valid neighboring intervals, were excluded rather than interpolated to avoid edge-related interpolation distortion. For each case, the number of detected R peaks, NN intervals analyzed, ectopic/artifact intervals, and intervals corrected by interpolation were documented. Only recordings with <5% ectopic/artifact contamination were retained for final HRV analysis. Per-case signal-quality metrics, raw RR tachograms, and corrected NN tachograms are provided in Supplementary Table 1, Supplementary Figure 1. The excluded patient with significant atrial dysrhythmia is shown separately in Supplementary Figure 2.

Time-domain metrics included mean RR interval, SDNN, RMSSD, and pNN50. Frequency-domain analysis included high-frequency (HF) power and the LF/HF ratio. Conventional frequency band limits were defined as 0.04–0.15 Hz for low frequency (LF) and 0.15–0.40 Hz for high frequency (HF). Prior to spectral analysis, RR interval series were interpolated and resampled at 4 Hz and linearly detrended. Power spectral density was estimated using Welch’s method with a maximum segment length of 256 samples and 50% overlap.

Nonlinear analysis included histogram-based Shannon entropy and two-dimensional Poincaré plot analysis. Shannon entropy was calculated from the probability distribution of RR intervals using a histogram-based approach with 20 bins. No phase-space embedding or delay reconstruction parameters were applied, as the entropy analysis was performed directly on the one-dimensional RR interval distribution. Two-dimensional Poincaré plots were constructed by plotting each RR interval (RRn) against the subsequent interval (RRn+1). From these plots, short-term variability (SD1, transverse axis) and long-term variability (SD2, longitudinal axis) were calculated using standard geometric formulations to characterize attractor morphology and variability dispersion.

HRV results were expressed both as absolute values and as exploratory standardized Z-scores relative to age- and sex-specific reference values reported by Voss et al. (2015). That study provides short-term HRV reference distributions from 5-minute ECG recordings obtained in healthy subjects from the KORA S4 cohort, stratified by sex and age categories. For each patient, the corresponding reference group was selected according to sex and age decade. Z-scores were calculated as: Z = (observed value − reference mean)/reference standard deviation.

The reference mean and standard deviation used for each metric and each case are reported in Supplementary Table 2. Z-score normalization was used only as a descriptive approach to contextualize individual HRV values against published healthy reference distributions and was not used for formal statistical inference or diagnostic classification. Because the reference values from Voss et al. were derived from 5-minute ECG recordings, whereas the present analysis used approximately 10-minute analyzable pre-HBOT segments, standardized deviations were interpreted cautiously.

Entropy-based Z-scores were interpreted with particular caution because entropy estimates are method-dependent. The reference entropy metric available in Voss et al. is not identical to the histogram-based Shannon entropy calculated in the present analysis; this distinction is detailed in Supplementary Table 2. Therefore, entropy Z-scores were used only as approximate exploratory comparators rather than as normative classifications.

Results

Eight male and female divers presented with focal neurological deficits consistent with neurological decompression sickness (NDCS), including monoplegia, hemiparesis, and vestibular dysfunction. Clinical severity varied from moderate to severe across the cohort based on composite scoring (Boussuges, PSI, and MEDSUBHYP).

Decompression exposure reconstruction using the Bühlmann ZHL-16 model yielded maximum Bühlmann gradient fractions ranging from 52% to 139%, with the highest value observed in Case 4. All subjects achieved complete symptom resolution following a single hyperbaric oxygen therapy (HBOT) session using the USN TT6 protocol.

Time-domain HRV metrics

Time-domain HRV metrics did not show a uniform response pattern across the cohort. SDNN and RMSSD varied substantially between cases, with elevated values in some patients and reduced or near-reference values in others. This supports marked interindividual heterogeneity rather than a single time-domain HRV phenotype in NDCS. Similarly, pNN50 remained low in Cases 4, 7, and 8, suggesting attenuation of short-term beat-to-beat variability in these cases (Figures 1, 2).

Absolute heart rate variability (HRV) metrics across eight cases of neurological decompression sickness (NDCS). The figure displays absolute values for time-domain, frequency-domain, and nonlinear/descriptive rhythm-structure metrics for each of the eight subjects prior to hyperbaric oxygen therapy. Gray shaded regions represent normative reference ranges for a healthy adult population (Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology, 1996). Time-domain metrics: SDNN (ms), RMSSD (ms), and pNN50 (%) demonstrate varied absolute variability across the cases, with some subjects exhibiting values significantly outside the normative gray bands. Frequency-domain metrics: HF Power (ms2) and the LF/HF ratio reflect the distribution of spectral power. Nonlinear metrics: Shannon Entropy.

Frequency-domain HRV metrics

Frequency-domain HRV metrics also showed substantial interindividual variability. HF power was highest in Case 1 and was also relatively high in Case 6, whereas Cases 4, 7, and 8 showed substantially lower HF power. LF/HF ratio values were lower than the corresponding age- and sex-specific reference values in several cases, particularly Cases 4 through 8. Because respiration was spontaneous and not monitored, HF power and LF/HF ratio were interpreted as descriptive frequency-domain indices rather than direct measures of parasympathetic or sympathovagal balance.

Nonlinear HRV dynamics

Among the nonlinear descriptors, Shannon entropy was lower in a subset of cases, most clearly in Cases 7 and 8. Thus, entropy-based findings suggested heterogeneous rather than uniformly reduced distributional complexity of the NN interval series. These findings were interpreted as exploratory because entropy estimates are sensitive to preprocessing, binning strategy, segment selection, and the reference method used for comparison. Two-dimensional Poincaré plots further illustrated the variability in attractor geometry across cases (Figure 3).

Two-dimensional Poincaré plots for cases 1–8. Each plot shows the relationship between consecutive RR intervals (RRn vs. RR), with the longitudinal and transverse axes representing SD2 and SD1, respectively. The plots illustrate the varying degrees of total variability and attractor geometry across the patient cohort prior to hyperbaric intervention.

Discussion

This case series represents, to our knowledge, one of the few available human descriptions of HRV in NDCS. Given the rarity and clinical heterogeneity of neurological decompression sickness, this study should be interpreted as an exploratory case series intended to generate physiological hypotheses rather than establish definitive autonomic patterns.

Based on these eight cases of neurological decompression sickness (NDCS), two principal observations emerge regarding autonomic dysfunction as assessed by HRV. First, Shannon entropy showed a heterogeneous but clinically relevant pattern, with near-reference values in some cases and clear reductions in others, particularly Cases 7 and 8. Therefore, entropy did not identify a uniform cohort-wide abnormality, but it remained useful