심호흡이 자율신경안정에 미치는 영향, 최신 논문 탐구!!

[문형철]

심호흡(느린 깊고 diaphragmatic/slow breathing)이

부교감신경을 활성화(주로 HRV 증가, vagal tone 강화, baroreflex 개선)하여

통증·스트레스·자율신경 관련 질환을 개선한다는 증거는

다수의 메타분석·RCT·체계적 고찰에서 잘 입증.

최신(2023~2026) 고영향력 논문(고인용·Cell/Nature family·전문 저널 등) 중

핵심적인 몇 가지를 선별해 메커니즘과 임상 효과 중심으로 정리

1. 최고 영향력: Breathwork(호흡 기법, cyclic sighing 포함) 메타분석 – Cell Reports Medicine (2023)

• Balban MY et al. (2023). Brief structured respiration practices enhance mood and reduce physiological arousal. Cell Reports Medicine, 4(1):100895. (인용 ~474회, Cell family 고영향력 저널)
• 주요 내용:
• 하루 5분 cyclic sighing(깊고 긴 호기 중심 심호흡 변형)을 비롯한 breathwork가 mindfulness meditation보다 기분 개선·생리적 각성 감소(호흡수↓) 효과가 더 크다는 RCT.
• 긴 호기가 부교감신경 활성화를 유도해 vagal tone을 높이고 스트레스/불안을 줄임.
• . “5분 호흡이 불안·기분에 미치는 효과”로 널리 인용됨.

연구 설계:

원격 무작위 대조 실험 (RCT, NCT05304000).

약 1개월 동안 매일 5분씩 호흡 연습(Breathwork) 또는 mindfulness meditation을 수행한 참가자들을 비교

주요 비교 그룹 (하루 5분씩)

1. Cyclic Sighing (순환 한숨): 길게 내쉬는 호흡 강조 (가장 효과적)
2. Box Breathing: 들이쉬기-참기-내쉬기-참기 균등
3. Cyclic Hyperventilation with Retention: 빠른 들이쉬기 + 짧은 내쉬기
4. 대조군: Mindfulness Meditation (명상)

주요 결과 (Breathwork 특히 Cyclic Sighing)

• 기분 개선: Breathwork 그룹이 meditation보다 유의미하게 더 큰 향상 (p < 0.05)
• 불안 감소 및 생리적 각성 감소: 호흡수, 심박수 등에서 개선
• 특히 Cyclic Sighing이 mood 향상과 호흡수 감소에서 가장 뛰어남
• 모든 호흡 연습과 명상 모두 일상 불안 감소와 기분 개선에 효과 있었으나, Breathwork가 더 우수

하루 5분 Cyclic Sighing 같은 간단한 구조화된 호흡 연습이

스트레스 관리, 기분 향상, 자율신경계 안정에 매우 효과적인 도구로 제시.

명상보다 접근하기 쉽고 즉각적인 효과가 강점

https://link.springer.com/article/10.1007/s12671-025-02660-2

마인드풀니스 명상(mindfulness meditation)의

건강 효과(스트레스 감소, 웰빙 향상 등)는 잘 알려져 있지만,

정확한 기전(mechanism)은 불분명합니다.

Paced breathing(의도적으로 호흡 속도를 늦추는 호흡 조절)도

비슷한 효과를 보이는데,

둘 다 호흡에 초점을 맞춘다는 공통점이 있습니다.

이 연구는

EEG(뇌파)와 SCL(피부 전도도, 교감신경 활성도 지표)를 통해

두 기법이 뇌 활동(신경 진동, neural oscillations)과 각성(arousal)에 미치는 차이를 비교했습니다.

연구 방법

• 참가자: 건강한 젊은 성인 80명 (평균 연령 20.15세, 여성 64명).
• 무작위 배정:
  • 마인드풀니스 명상 그룹 (n=40): 자연스러운 호흡 관찰, 비판단적 주의.
  • Paced breathing 그룹 (n=40): 의도적 느린 호흡 조절.
• 측정: 과제 전 5분 휴식(pre-task), 과제 중(task), 과제 후 5분 휴식(post-task) 동안 EEG(전 주파수대)와 SCL 기록.

1. Mindfulness Meditation (왼쪽 열)

• Pre(파랑) → Task(빨강) → Post(초록) 변화가 전반적으로 매우 작음.
• 가장 두드러진 변화: Alpha 대역(약 8~12 Hz)에서 Task 동안 진폭이 약간 감소.
• 다른 주파수대(delta, theta, beta, gamma)에서는 거의 변화 없음.
• Task 후(Post)에는 거의 Pre 상태로 돌아감.
• 결과: 명상 중 alpha power가 감소 → 주의 집중이나 내면 지향 상태를 반영.

2. Paced Breathing (오른쪽 열)

• Task(빨강) 동안 전반적인 진폭이 증가하는 경향이 뚜렷함.
• 특히 저주파 대역(delta ~ theta)과 alpha에서 증가가 더 명확.
• Task 후(Post)에도 일부 증가가 남아 있지만, Pre에 가까워짐.
• 결과: 의도적 느린 호흡이 뇌 전체의 neural oscillations(신경 진동)을 강화.

주요 결과

• Paced breathing (의도적 호흡 조절):
  • 모든 EEG 주파수대(delta, theta, alpha, beta, gamma)에서 전반적 진폭 증가.
  • SCL(교감신경 활성) 증가 → 더 높은 각성 상태.
  • 호흡 속도 크게 감소.
• Mindfulness meditation (마인드풀니스 명상):
  • Alpha 진폭 감소 (다른 주파수대는 큰 변화 없음).
  • SCL은 거의 변화 없음 (각성 유지).
  • 호흡 속도는 약간 감소하지만 paced breathing만큼 크지 않음.
• Alpha와 SCL 관계: 전반적으로 유의미한 상관관계 없음 (기존 연구와 다름).
• 과제 후(post-task) 효과는 지속되지 않음 (즉각적 상태 변화 중심).

결론 및 함의

• 두 기법은 서로 다른 기전을 가진다:
  • Paced breathing: 호흡 조절을 통해 자율신경계 직접 영향 → 뇌파 전반 증가 + 교감 활성 변화.
  • Mindfulness meditation: 주의(attention) 중심 → alpha 감소 (내면 주의나 억제 관련) + 각성 유지.
• 건강 효과(예: 스트레스 감소)는 공통적이지만, underlying process는 다르다는 점을 시사. 이는 명상과 호흡 기법을 상황에 맞게 선택하는 데 도움이 될 수 있습니다. (예: 이완 목적 vs. 주의력 향상 목적)

2. 최신 체계적 고찰: Diaphragmatic Breathing(심횡격막 호흡) 건강 효과 – Complementary Therapies in Medicine (2026)

• Kwon CY et al. (2026). The health effects of diaphragmatic breathing: A systematic review. Complementary Therapies in Medicine, 96:103317. (48개 RCT 종합)
• 주요 내용:
• '심횡격막 호흡이 자율신경계(ANS) 조절(부교감 우위, sympathetic↓) 통해 GERD(위식도역류, 약물 사용↓), 불안, post-COVID 증후군, 임신당뇨 등에서 일관된 개선 효과 확인.
• '건강인에서는 급성 심혈관 이점도 있음. COPD나 심장수술 후에는 효과가 불일치하거나 표준치료보다 약함.

주요 방법 및 특징

• 횡격막 호흡(Diaphragmatic Breathing, DB): 복식호흡, 복부호흡 등으로 불림
• 프로토콜이 매우 다양: 호흡수(2~10회/분), 세션 시간(3~45분), 기간(단회 ~ 12주)
• 방법론적 질: 매우 낮음 (결과 지표 중 2.12%만 low risk of bias)

주요 결과 (일관된 긍정 효과)

• 위식도역류질환 (GERD): 가장 강력한 증거 — 증상 개선 + 약물 사용 감소
• 불안(anxiety): 일관된 개선
• 코로나19 후유증 (post-COVID-19 syndrome): 증상 완화
• 임신성 당뇨 (gestational diabetes): 긍정적 효과
• 건강한 성인: 급성 심혈관계 이점 (혈압·심박수 등 단기 개선)

3. Breathwork 전체 메타분석 – Scientific Reports (Nature, 2023)

• Fincham GW et al. (2023). Effect of breathwork on stress and mental health: A meta-analysis of randomised-controlled trials. Scientific Reports. (인용 ~191회, 접근 35만 회 이상)
• 주요 내용:
• 느린 페이스 호흡(slow-paced, 5~6회/분)이 HRV 증가·부교감신경 활성화를 통해 스트레스(g=-0.35), 불안(g=-0.32), 우울(g=-0.40) 증상을 유의하게 줄임.
• 비임상군에서 특히 효과적.
• 메커니즘: Polyvagal theory 등으로 설명되며, slow breathing이 vagal tone을 높여 정신건강 개선.

• 연구 유형: 메타분석 (Meta-analysis of RCTs)
  • 검색 기간: 2022년 2월까지 (PsycInfo, PubMed 등 7개 DB)
  • 초기 검색: 1,325건 → 최종 포함: Stress 12편 (n=785), Anxiety 20편, Depression 18편

주요 결과

• 주요 결과 (Stress): 12개 RCT
  • Breathwork가 대조군 대비 스트레스 유의하게 감소 (Hedges’ g = −0.35, 95% CI −0.55 ~ −0.14, p = 0.0009)
  • 효과 크기: 작음~중간 (small-to-medium)
• 부차적 결과:
  • 불안 (Anxiety): g = −0.32 (p < 0.0001)
  • 우울 (Depression): g = −0.40 (p < 0.0001)
  • 모두 작음~중간 효과 크기, 통계적으로 유의함

4. 최신 RCT: Long COVID에서 Slow-Paced Breathing – American Journal of Medicine (2025)

• Mauro M et al. (2025). Heart Rate Variability Modulation Through Slow-Paced Breathing in Health Care Workers with Long COVID: A Case-Control Study. The American Journal of Medicine, 138(5):870-883.e5.
• 주요 내용:
• Long COVID 환자에서 slow-paced breathing(SPB, 6회/분)이 vagal response 강화·baroreflex sensitivity 개선 (HRV 지표 상승).
• 자율신경 불균형(교감 우위)을 직접적으로 교정.
• 임상 의미:
• Long COVID 같은 만성 염증·자율신경 질환에 적용 가능.
• 이전 연구와 함께 염증 마커 감소(예: IL-6, CRP)와 연계됨.

연구 방법

• 참가자: 총 126명 의료진 (HCWs)
  • G1: Long COVID (n=58) — 감염 후 평균 419.5일 경과, 증상 지속
  • G2: Never infected (n=53) — SARS-CoV-2 음성
  • G3: Fully recovered (n=14) — 완전 회복
• 평가:
  • 설문지: 정신신체 증상 (psychosomatic symptoms)
  • Finger Photoplethysmography로 HRV 측정 (자연 호흡 vs Slow-Paced Breathing)
• Slow-Paced Breathing (SPB): vagal tone 자극, 약 6회/분 (0.1 Hz resonance frequency) 호흡

주요 결과

• Long COVID 그룹 특징:
  • 정신신체 증상 유의하게 많음
  • 자연 호흡 시 0.1 Hz Mayer wave (혈압 진동, baroreflex arc) 검출률 낮음 (19% vs 대조군 42% / 40%, p=0.027) → 혈관 조절 및 baroreflex 기능 저하
• Slow-Paced Breathing 효과:
  • SPB 동안 Mayer wave 검출률이 대조군 수준으로 회복 (91.2% vs 100% / 100%, p=0.09)
  • 다른 HRV 지표들은 그룹 간 큰 차이 없음
  • 결론: Long COVID에서 resting 상태 baroreflex/vascular modulation이 손상되었으나, SPB로 효과적으로 개선됨

결론 및 임상적 의미

• Long COVID 환자에서 자율신경 불균형(특히 baroreflex impairment)이 관찰됨
• 느린 호흡(Slow-Paced Breathing)은 비침습적·저비용 방법으로 vagal 자극과 baroreflex 회복에 효과적
• 반복적인 호흡 연습이 장기적으로 혈관 반응성 회복과 심혈관 위험 감소에 도움이 될 수 있음

5. 통증 관련 최근 증거 (2024)

• Rezaei M et al. (2024). Effect of deep diaphragmatic breathing on pain in patients with cancer. Heliyon (또는 PMC11609224). RCT로 암 환자에서 심횡격막 호흡이 통증 지각 유의하게 감소.
• 추가로 Joseph AE et al. (2022) 메타분석 등에서 slow deep breathing이 급성·만성 통증 완화에 효과적이며, 부교감 활성화(교감↓, HRV↑)가 핵심 기전으로 지목됨.

• 목적: 전이성 위장관암(metastatic gastrointestinal cancers) 환자에서 Deep Diaphragmatic Breathing (깊은 횡격막 호흡)이 급성 및 만성 통증에 미치는 효과 확인

연구 방법

• 참가자: 44명 (종양병동 입원 환자, 편의 표본 추출 후 무작위 배정)
  • 실험군 (Case): 횡격막 호흡 중재
  • 대조군 (Control): 일반 간호만
• 중재: 10일 동안 하루 2회, 각 10분 Deep Diaphragmatic Breathing
• 평가 도구: 통증 평가 도구 (Acute pain & Chronic pain) — 중재 전후 측정
• 분석: SPSS 24 사용

주요 결과

• 연령: 평균 53.95 ± 10.51세
• 급성 통증 (Acute Pain):
  • 중재 전: 실험군 3.50 ± 1.84 vs 대조군 2.18 ± 1.65 (p=0.01)
  • 중재 후: 실험군 1.72 ± 1.07 (감소) vs 대조군 3.72 ± 1.95 (증가) → p=0.001 (유의미한 차이)
• 만성 통증 (Chronic Pain):
  • 중재 전: 두 군 간 유의 차이 없음 (p=0.07)
  • 중재 후: 실험군에서 감소, 대조군에서 증가 → p=0.01 (유의미한 차이)

결과 해석: 10일간의 짧은 중재만으로도 급성·만성 통증 모두 유의하게 감소 (대조군은 악화 경향)

결론 및 임상적 의미 (저자)

• Deep Diaphragmatic Breathing은 전이성 위장관암 환자의 통증 관리에 효과적인 비약물적 중재
• 저비용, 안전, 쉽게 적용 가능 → 일상 간호 프로그램에 포함 추천
• “알려지지 않은 요인에도 불구하고 급성·만성 통증 조절에 도움이 된다”고 강조

요약 및 실천 팁

• 강력한 메커니즘 증거: Laborde et al. (2022, Neurosci Biobehav Rev) 메타분석(223개 연구)에서 voluntary slow breathing(5~7회/분)이 vagally-mediated HRV를 가장 강력하게 증가시킨다는 결론.
• 효과적인 프로토콜: 6회/분(호기 길게, 1:2 비율), 하루 5~20분. Cyclic sighing이나 diaphragmatic breathing이 가장 실증됨.

Differential Effects of Mindfulness Meditation and Paced Breathing on Neural Oscillations and Arousal

• ORIGINAL PAPER
• Open access
• Published: 15 August 2025

• Volume 16, pages 2683–2698, (2025)
• Cite this article

You have full access to thisopen accessarticle

Download PDF

Save article

MindfulnessAims and scopeSubmit manuscript

Differential Effects of Mindfulness Meditation and Paced Breathing on Neural Oscillations and Arousal

Download PDF

• Alexander T. Duda, 
• Adam R. Clarke & 
• Robert J. Barry 

• 3483 Accesses

• 3 Citations

• Explore all metrics 

Abstract

Objective

Although research continues to explore the broad applications of mindfulness meditation, the precise mechanisms underlying its health benefits remain unclear. Paced breathing, which involves intentionally slowing the breath, shows similar benefits. The shared focus on breath in both practices may partly explain their positive health effects. This study explored the impact of mindfulness meditation and paced breathing on electroencephalography (EEG) and skin conductance level (SCL), to investigate changes in brain function and sympathetic arousal.

Method

Eighty healthy young adults were randomly allocated to either a mindfulness meditation (n = 40; Mage = 20.15; 32 females) or paced breathing (n = 40; Mage = 20.15; 32 females) task. EEG and SCL data were recorded during the allocated task and two 5 min eyes-closed resting states, immediately before and after the task. The effects of condition (mindfulness meditation vs. paced breathing) and state (pre-task vs. task; pre-task vs. post-task) were investigated.

Results

Significant interactions between condition and state (pre-task vs. task) demonstrated increased global EEG amplitudes across all frequency bands, as well as increased SCL, during paced breathing. In contrast, mindfulness meditation resulted in a decrease in global alpha amplitude, while the remaining frequency bands and SCL remained somewhat consistent when comparing pre-task with task. No association was found between global alpha amplitude and SCL.

Conclusions

These findings highlight the differing effects of mindfulness meditation and paced breathing on neural oscillations and arousal, suggesting that each practice involves unique underlying mechanisms that contribute to their respective health benefits.

Preregistration

This study was not preregistered.

Similar content being viewed by others

The Effect of Slow-Paced Breathing on Cardiovascular and Emotion Functions: A Meta-Analysis and Systematic Review

Article Open access02 January 2024

How does paced breathing influence post-emotion encoding?

Article Open access10 April 2026

Self-Directed Versus Pre-Determined Mindfulness Training: A Longitudinal Study on Perceived Stress, Well-Being, and Mindfulness

Article Open access28 March 2026

Explore related subjectsDiscover the latest articles, books and news in related subjects, suggested using machine learning.

• Brain Mapping
• Brain Stimulation
• Neurophysiology
• Physiological Psychology
• Slow-wave sleep
• Electroencephalography
• Mindfulness-Based Interventions for Mental Health

With origins dating back centuries, mindfulness has gained significant popularity over the past 50 years, particularly since its adaptation for stress reduction in therapeutic and clinical settings in the 1970s (Van Dam et al., 2018). It has since expanded into diverse environments, with mindfulness interventions widely accessible through courses, workshops, therapy, online platforms, and smartphone applications (Birtwell et al., 2019; Crane et al., 2017). While research strives to keep up with its broad applications, the exact mechanisms behind mindfulness’s many reported psychological and physical health benefits remain unclear (Creswell, 2017; Hölzel et al., 2011; Ludwig & Kabat-Zinn, 2008). Mindfulness is commonly defined as the awareness arising from purposefully directing non-judgmental attention to present-moment experiences (Kabat-Zinn, 1994). Mindfulness can be cultivated through various practices, one of which is mindfulness meditation, which typically involves practitioners sitting quietly with eyes closed and focusing attention on an anchor, such as the breath (Birtwell et al., 2019; Kabat-Zinn, 2003).

Mindfulness meditation is associated with lasting changes in brain structure and function, particularly in areas involved in emotional regulation, attention, and self-awareness (Tang et al., 2015). Changes in neural oscillations linked to mindfulness meditation have been measured through electroencephalography (EEG), and although there is no consensus on the specific effects, emerging patterns link certain frequency bands with distinct cognitive processes and functions (Lee et al., 2018; Lomas et al., 2015). Studies examining mindfulness meditation typically investigate state changes, which are the immediate effects that occur during meditation, and trait changes, which are the lasting effects that persist outside of meditation (Cahn & Polich, 2006). The repeated practice of mindfulness meditation is believed to underpin trait changes, as the skills learnt during meditation extend into everyday life, resulting in lasting neuroplastic changes in the brain and associated health benefits (Davidson & Kaszniak, 2015; Kiken et al., 2015; Tang et al., 2015).

Research has shown that different styles of meditation elicit varying impacts on neural oscillations, both regionally and globally, which also vary depending on the practitioner's expertise (Cahn & Polich, 2006; Lee et al., 2018; Lomas et al., 2015; Tang et al., 2015). This variability makes it challenging to generalise findings and identify the precise mechanisms of change. Consequently, the state changes induced by meditation remain poorly understood, and research findings are somewhat inconsistent.

Despite variations in the types of meditation examined and some methodological inconsistencies, a substantial body of EEG research has reported increases in theta (~ 4.0–7.5 Hz) and alpha (~ 8.0–13.0 Hz) oscillations during meditation compared to rest (Cahn & Polich, 2006; Lee et al., 2018; Lomas et al., 2015). These oscillatory patterns have been suggested to characterise meditation as a state of relaxed alertness (Britton et al., 2014; Lomas et al., 2015) and align with theoretical accounts of mindfulness meditation that emphasise the central role of attention (Hölzel et al., 2011; Lutz et al., 2008). Additionally, these findings are consistent with the broader literature linking shifts in alpha oscillations to attention and inhibition processes (Cooper et al., 2003; Klimesch, 2012; Klimesch et al., 2007; Knyazev, 2007). Theta oscillations have similarly been associated with attention processes (Karakaş, 2020; Takahashi et al., 2005) as well as emotion regulation (Knyazev, 2007).

Comparatively, the remaining frequency bands, delta (~ 1.0 to 3.5 Hz), beta (~ 13.5 to 30.0 Hz), and gamma (~ 30 + Hz), have been investigated less consistently. Tentative evidence suggests that increases in delta and beta oscillations during meditation may be linked to attentional processes (Lee et al., 2018; Lomas et al., 2015). Increases in gamma oscillations may reflect an altered state of consciousness, particularly among experienced practitioners, although research on gamma oscillations in meditation is relatively new compared to other frequency bands (Lee et al., 2018).

Collectively, these results suggest that attention is a major contributing factor to the neural changes induced by mindfulness meditation. However, investigating the association between attention and neural processes is not always straightforward. Cooper et al. (2003) found that alpha amplitude was greater during internally directed attention compared with externally directed attention, regardless of sensory modality, highlighting the complexity of interpreting these patterns. Such findings suggest that the type of attentional focus may interact with inhibitory processes in ways that complicate interpretation. These nuances may help account for inconsistencies in the EEG meditation literature, where conflicting results are not unexpected given the variability in attentional demands across different practices. To better understand the unique effects of mindfulness meditation, it is essential to distinguish the role of attentional focus from that of breath regulation. The benefits of mindfulness meditation may, in part, arise from its distinct impact on breathing function, potentially setting it apart from other forms of meditation.

Breathing can be divided into two categories: spontaneous breathing, where no effort is made to alter breath rate, and paced breathing, where breath rate is deliberately altered (Park & Park, 2012). From a physiological perspective, slow or paced breathing (typically 0.1 Hz) regulates the autonomic nervous system via vagus nerve stimulation and respiratory influence on the cardiovascular system, leading to changes in neural oscillations that modulate arousal and likely underpin the positive physiological and psychological health effects e.g., improved well-being, cognitive enhancement, reduced stress, anxiety, and depression (Balban et al., 2023; Brown et al., 2013; Epe et al., 2021; Gerritsen & Band, 2018; Jerath et al., 2015; Mourya et al., 2009; Wielgosz et al., 2016; Zaccaro et al., 2018, 2022). Unlike paced breathing, mindfulness meditation is rarely concerned with altering the breath, but instead emphasises observing the natural, spontaneous breath as it unfolds (Brenner et al., 2020). However, during mindfulness meditation, breathing rate tends to slow on its own, and research has identified that more experienced practitioners demonstrate greater rate decreases (Wielgosz et al., 2016).

Balban et al. (2023) compared the effects of short daily breathwork with mindfulness meditation, finding that both improve mood and reduce anxiety, but breathwork improved mood and reduced physiological arousal more than mindfulness meditation. EEG studies have also investigated paced breathing, with mixed findings. Stancák et al. (1993) found that paced breathing at 0.1 Hz decreased alpha power compared to rest, particularly at parietal and occipital electrodes, while beta power was significantly increased during paced breathing. In comparison, Fumoto et al. (2004), Yu et al. (2011), and Park and Park (2012) found increased alpha and decreased theta power during slow breathing, however analyses and experimental design, particularly the electrodes of interest, differed between these studies.

Considering that the health benefits of both mindfulness meditation and paced breathing are likely mediated by attention and autonomic nervous system regulation (Amihai & Kozhevnikov, 2015; Zaccaro et al., 2018), arousal may be a key mechanism underpinning these effects. Arousal, a person’s level of alertness and attentiveness, can be indexed through skin conductance level (SCL), the "gold standard" measure of sympathetic nervous system activity (Barry et al., 2005; Hicks et al., 2020). Utilising skin conductance, Scavone et al. (2020) demonstrated that a mindfulness breathing exercise reduced arousal during a stressful task, with greater reductions among individuals higher in state mindfulness. Similarly, slow paced breathing has been shown to buffer against stress responses, evidenced by reduced SCL during exposure to physical and psychological stressors (Harris et al., 1976). Supporting these findings, Hicks et al. (2020) found that higher trait mindfulness and lower SCL were associated with less perceived stress among undergraduate students, further linking arousal regulation with mindfulness.

Importantly, research has also identified a negative relationship between SCL and global alpha oscillations, with increased alpha associated with reduced SCL during resting states (Barry et al., 2005, 2020). This suggests that global alpha may serve as an inverse index of arousal. In this context, reports of increased alpha activity during mindfulness meditation and paced breathing may indicate a reduction in arousal. However, not all meditative states produce the same neural and autonomic changes. Travis and Wallace (1999) found that transcendental meditation reduced SCL and breathing rate without altering alpha activity. Additionally, Amihai and Kozhevnikov (2014) found distinct effects on EEG and the autonomic nervous system across four different types of meditation. These findings suggest that neural activity and physiological responses may vary according to the type of meditation.

Given their shared benefits, it is important to distinguish between the mechanisms underlying mindfulness meditation and paced breathing. Both practices influence attention and autonomic regulation, suggesting arousal modulation as a shared mechanism. However, mindfulness meditation involves passive, non-judgmental awareness of spontaneous breathing, whereas paced breathing requires deliberate, active breath control. This distinction provides an opportunity to examine how neural oscillations and arousal changes relate to mindfulness dimensions versus intentional breath regulation.

The present study aimed to disentangle the state-specific effects of mindfulness meditation and paced breathing by examining changes in global EEG oscillations and sympathetic arousal. Building on recent work using a comparable mindfulness meditation (Duda et al., 2024a, 2024b, 2025), we predict that mindfulness meditation will be associated with a reduction in alpha oscillations and a stable or slight reduction in SCL. Such findings would suggest that attentional processes during mindfulness meditation operate independently of deliberate autonomic regulation and are consistent with characterising this state as one of relaxed alertness. In contrast, we expect that paced breathing will be associated with an increase in alpha oscillations and a reduction in SCL, as slow, controlled breathing directly engages autonomic regulation, promoting relaxation and decreasing sympathetic activation. This pattern of changes would support the role of arousal modulation as a central mechanism underlying the health benefits of paced breathing. Additionally, we predict increases in theta oscillations during both mindfulness meditation and paced breathing, consistent with attentional engagement, as well as increases in delta and beta oscillations, assuming these bands index attentional processes. Differences between conditions are expected to reflect subtle distinctions in cognitive and attentional demands. Finally, given the cognitive complexity and variability inherent in mindfulness meditation, particularly for novice meditators, we do not anticipate consistent changes in gamma oscillations. However, such activity may emerge more prominently during paced breathing, given its relative cognitive simplicity.

Method

Power Analysis

An a priori power analysis using G*Power (Faul et al., 2007) for a repeated-measures multivariate analysis of variance (MANOVA) (within-between interaction) indicated that, to detect a large effect size (f(V) = 0.4, equivalent to ηp2 = 0.14), an alpha level of 0.05, and a power of 0.80, a total of 64 participants were needed for two groups and three measurements.

Additionally, for the relationship between alpha and SCL for a two-tailed test with a correlation coefficient (ρ) of 0.32 (Barry et al., 2020), an alpha level of 0.05, and a power of 0.80, a total sample size of 74 participants was required.

Participants

A total of 80 participants, aged 18 to 35 years, were primarily recruited through the university's psychology research participation scheme, receiving course credit for their involvement. Participants were split randomly into two groups: mindfulness meditation (n = 40; M = 20.15, SD = 2.90 years; 32 females) or paced breathing (n = 40; M = 20.15 years, SD = 3.06; 32 females). Participants had varying levels of lifetime meditation experience (M = 31.73, SD = 70.46, range: 0–370 h), with the majority beginners (n = 74) and the remainder intermediate (n = 6), based on criteria by Van Dam et al. (2024).

Participants were non-smokers, fluent in English, and either right-handed (n = 77) or mixed-handed (n = 3), as assessed by the Edinburgh Handedness Inventory – Short Form (Veale, 2014). They were not taking any psychoactive medications or substances and had no history of epilepsy, cardiovascular, or neurological diseases, serious head injury, or periods of unconsciousness. Two participants reported a history of asthma but indicated that they were not experiencing any issues during the experimental session, while the remaining participants had no underlying respiratory issues. Participants were advised to abstain from alcohol, caffeine, and tobacco for at least 4 h prior to the experimental session, with all confirm‎ing adherence upon attendance. They were also requested to abstain from food and drink for at least 2 h before the experiment, except for small sips of water. Additionally, participants were instructed to follow a normal sleep routine and avoid intense or strenuous physical activity on the day of the experiment.

Electrophysiological Recording

EEG data were recorded using tin electrodes in a 19-channel electrode cap (Fp1, Fp2, F7, F3, Fz, F4, F8, T7, C3, Cz, C4, T8, P7, P3, Pz, P4, P8, O1, O2) and A2, with all referenced to A1, and grounded by an electrode placed between Fpz and Fz, in accordance with the International 10/20 System (Pivik et al., 1993). Electrooculogram (EOG) activity was recorded from four tin electrodes positioned above and below the right eye to measure vertical eye movements, and on the left and right outer canthi to measure horizontal eye movements. Electrocardiographic (ECG) data were recorded from two Ag/AgCl electrodes placed on the left and right forearms. EEG, EOG, and ECG data were recorded at 500 Hz, DC-70 Hz through a Neuroscan SynAmps2 amplifier and Acquire Software (Compumedics, Version 4.5.1) with electrode impedance levels kept below 5 kΩ. SCL was recorded from two sintered Ag/AgCl electrodes taped to the distal phalanges of the middle and index fingers of the left (non-dominant) hand. An electrolyte of 0.05 M NaCl in an inert ointment base was used, and SCL data were concurrently sampled at 500 Hz from a constant-voltage device (UFI BioDerm 2701) set at 0.5 V.

Procedure

After providing informed consent, participants were screened for eligibility. Eligible participants completed several brief questionnaires on demographics, mindfulness, and mental health (not reported here). Electrophysiological recording equipment was then fitted, and participants were seated in a dimly lit, air-conditioned, sound attenuated room. Seated upright in front of a 17-inch monitor, they first completed a brief EOG calibration task (Croft & Barry, 2000), followed by 5 min of resting with eyes-closed and 5 min with eyes-open, in counterbalanced order. Participants were then randomly assigned to either a 10 min mindfulness meditation or paced breathing task, delivered through headphones (Sony MDR-V700). After the task, an additional 5 min of resting with eyes-closed and 5 min with eyes-open (counter-balanced) were recorded. Given the differences between eyes-closed and eyes-open resting conditions, Barry et al. (2007) recommend using eyes-closed as a baseline in studies that do not involve eyes-open conditions or visual stimuli. Accordingly, only data from the task and eyes-closed resting state are reported here.

Mindfulness Meditation

The mindfulness meditation task was adapted from Kabat-Zinn (2006) Mindfulness of Breathing, designed to introduce beginners to meditation. This meditation focuses on attending to the breath, while purposefully directing non-judgemental awareness to the present moment. Participants received a brief overview of the task and had the opportunity to ask questions before starting. They were then instructed to sit comfortably, close their eyes, and begin focusing on their breath. The session began with 5 min of guided audio instruction, followed by 5 min of silence during which participants continued the practice on their own, for a total duration of 10 min.

Paced Breathing

The paced breathing task started with a rate of approximately 10 breaths/min (0.17 Hz), with 3 s for each inhalation and exhalation, for the first 2 min. The respiratory rate was then further slowed to approximately 6 breaths/min (0.1 Hz), with 5 s for each inhalation and exhalation, for the remaining duration. Participants were given a brief overview of the paced breathing task and had the opportunity to ask questions prior to starting. They were then instructed to sit comfortably, close their eyes, and slow their breathing. An audio count guided participants through the first 5 min, after which they continued the task in silence for the remaining 5 min at 6 breaths/min, totalling 10 min.

Data Processing and Quantification

All EEG and SCL data processing was completed in MATLAB (R2023b) using EEGLAB (Delorme & Makeig, 2004; v2023.0) and custom scripts. The EEG data were first corrected for ocular artefact using the Revised Aligned-Artefact Average Procedure (RAAA; Croft & Barry, 2000) and then digitally re-referenced to mean ears (A1, A2). Following visual inspection, bad channels (excessive noise/artefact) were interpolated using spherical spline interpolation (M = 4.8%, SD = 5.3%; range: 0–3 channels). One hundred and fifty sequential 2 s epochs, each baselined across their duration, were extracted separately from each resting eyes-closed state (pre-task and post-task). Additionally, 150 sequential 2 s epochs, each baselined across their duration, were extracted from the mindfulness meditation and paced breathing tasks corresponding with the period of silence (task). Epochs with extreme amplitudes (± 300 µV), voltage jumps (> 50 µV between adjacent data points), or flat lines (less than 0.5 µV change within 100 ms intervals) for any channel were rejected. A 10% Hanning window was applied to the remaining epochs, followed by a Discrete Fourier Transformation (DFT) to derive single-sided amplitude spectra at 0.5 Hz frequency resolution, with a correction applied to compensate for the windowing effect.

EEG Band Quantification

Mean spectra at 0.5 Hz frequency resolution were calculated for each participant by averaging across epochs in each state (pre-task, task, post-task). Absolute amplitudes in the delta (1.0–3.5 Hz), theta (4.0–7.5 Hz), alpha (8.0–13.0 Hz), beta (13.5–30.0 Hz), and gamma (30.5–45.0 Hz) bands were summed at each electrode. These values were then averaged across the 19 scalp electrodes to obtain a global amplitude for each participant in each state for each frequency band.

SCL

SCL was measured in microSiemens (μS) and as with the EEG data, 150 sequential 2 s epochs were extracted separately from each eyes-closed resting state (pre-task and post-task), as well as 150 sequential 2 s epochs from mindfulness meditation and paced breathing task corresponding with the period of silence (task). SCL data were visually inspected for artefact and the mean levels for each participant in each state (pre-task, task, post-task) were calculated.

Respiratory Rate

ECG data were processed using Kubios HRV Premium (Version 3.5.0) to estimate respiratory rate. Series interpolation was performed at 4 Hz, and interval detrending was applied using the smoothness priors method with a smoothing parameter of 500 and a cutoff frequency of 0.035 Hz. Automatic beat correction was applied, and all corrected beats were visually inspected to verify detection accuracy and ensure artifacts were appropriately addressed (M = 0.2%, SD = 0.5% of total beats corrected). Respiratory rate was derived using the software’s respiratory rate estimation algorithm (Lipponen & Tarvainen, 2022), which has demonstrated strong correlations with true respiration during resting-state recordings (R = 0.892). The analysed time periods matched those used for EEG and SCL data. Respiratory frequency (Hz) was converted to breaths per minute (bpm; 1 Hz = 60 bpm).

Data Analyses

All analyses were completed in IBM Statistical Package for the Social Sciences (version 28). Data were checked for normality and outliers. Univariate outliers, defined as any value with a z-score of ± 2.58 (p = 0.01), were adjusted to one unit larger or smaller than the next closest value to ensure they remained the most extreme (Tabachnick & Fidell, 2019). In total, 3.7% of the EEG data, 2.9% of the SCL data, and 1.7% of the respiratory rate data were adjusted for outliers. The data were analysed using a repeated measures MANOVA with planned non-orthogonal contrasts to examine the within-subjects factor of state (pre-task vs. task; pre-task vs. post-task) with the between-subjects factor of condition (mindfulness meditation [MM] vs. paced breathing [PB]). An α of 0.05 was used to determine significance for all effects.

Due to the presence of outliers in the EEG and SCL data, correlations were run utilising Spearman’s rank-order correlation to investigate any relationship between SCL and global alpha amplitude for each condition in each state, as well as the difference between states (task minus pre-task; post-task minus pre-task). As these analyses were primarily exploratory, tests were two-tailed and eval‎uated at an α of 0.05.

Results

Respiratory Rate

As shown in Fig. 1, respiratory rate was significantly reduced during the task compared with pre-task, and this effect interacted significantly with condition, with a larger reduction observed in the paced breathing condition (see supporting statistics in Table 1). There was no significant change in respiratory rate between pre-task and post-task, and this effect did not interact with condition.

Fig. 1

The alternative text for this image may have been generated using AI.

Full size image

Mean respiratory rate (breaths per minute; bpm) for Mindfulness Meditation (MM) and Paced Breathing (PB) conditions in each State (Pre-task [Pre], Task, and Post-task [Post]). Standard error of the mean is represented in the error bars. * p < 0.05

Table 1 Statistical results for respiratory rate comparing Pre-task (Pre) vs. Task and Pre-task (Pre) vs. Post-task (Post), including interaction effects of Mindfulness Meditation (MM) vs. Paced Breathing (PB) conditions

Full size table

EEG

In the mindfulness meditation condition, the number of accepted epochs per participant ranged between 132–150 (M = 146.7, SD = 4.6) for pre-task, 132–150 (M = 146.5, SD = 4.6) for task, and 126–150 (M = 145.9, SD = 6.5) for post-task. In the paced breathing condition, the number of accepted epochs per participant ranged between 103–150 (M = 145.9, SD = 8.4) for pre-task, 130–150 (M = 144.9, SD = 6.0) for task, and 90–150 (M = 144.4, SD = 10.2) for post-task.

Figure 2 displays the grand mean spectral amplitudes for both conditions in each state. The mindfulness meditation data are shown on the left, with midline sites (Fz, Cz, Pz, O1/O2) and frequency bands labelled at O1/O2. As expected, there is a peak in the delta band, followed by a decrease in amplitude with increasing frequency until the typical increase in amplitude evident in the alpha band, which increases in size from the frontal to the occipital sites. The spectral amplitudes are similar both pre-task and post-task across all sites; however, there is a notable decrease in alpha amplitude during task in the mindfulness meditation condition, most prominent at O1/O2. In comparison, the paced breathing condition exhibits a similar overall pattern with a peak in the delta band, followed by a decrease in amplitude with increasing frequency, and a small increase in alpha amplitude. The pre-task and post-task spectra are quite similar except in O1/O2, where the pre-task alpha band peak is more prominent. Additionally, in task, there is a shift in peak frequency towards the theta band, as well as an increase in amplitude in the theta and alpha band evident at Fz, Cz, and Pz.

Fig. 2

The alternative text for this image may have been generated using AI.

Full size image

Grand mean spectral amplitude at Fz, Cz, Pz, and O1/O2 are displayed for the Mindfulness Meditation (left) and Paced Breathing (right) conditions in each State (Pre-task [Pre], Task, and Post-task [Post]). The EEG band ranges are indicated at O1/O2 on the Mindfulness Meditation plot. δ = delta (1.0–3.5 Hz); θ = theta (4.0–7.5 Hz); α = alpha (8.0–13.0 Hz); β = beta (13. 5–30.0 Hz); γ = gamma (30.5–45.0 Hz)

Topographically, delta appeared to be central and maximal at the midline in both conditions for all states (Fig. 3). Global delta amplitude increased significantly during the task compared with the pre-task, and this interacted significantly with condition, with a small increase during task in the mindfulness meditation condition and a large increase in the paced breathing condition which appeared maximal centrally, as shown in Fig. 4 (refer to Table 2 for supporting statistics). There was also a significant increase post-task relative to pre-task, which did not interact with condition.

Fig. 3

The alternative text for this image may have been generated using AI.

Full size image

Delta band (1.0–3.5 Hz) topographic headmaps (left) and global mean amplitude (right) for Mindfulness Meditation and Paced Breathing conditions in each State (Pre-task [Pre], Task, Post-task [Post]). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Fig. 4

The alternative text for this image may have been generated using AI.

Full size image

Topographic headmaps (left) and global mean amplitude (right) depicting the differences between pre-task and task (Task − Pre) and pre-task and post-task (Post − Pre) for Mindfulness Meditation (MM) and Paced Breathing (PB) conditions in the delta band (1.0–3.5 Hz). Δ = change in global mean amplitude (µV). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Table 2 Statistical results of EEG frequency bands comparing Pre-task vs. Task, and Pre-task vs. Post-task, including interaction effects of Mindfulness Meditation vs. Paced Breathing conditions

Full size table

Theta appeared maximal at the midline in all states and both conditions (Fig. 5). Global theta amplitude increased significantly during task compared with pre-task (see Table 2 for supporting statistics), and this interacted significantly with condition, with a small increase in mindfulness meditation, and a large increase in paced breathing which appeared maximal at midline sites, as shown in Fig. 6. There was also a significant increase post-task relative to pre-task, that did not interact with condition.

Fig. 5

The alternative text for this image may have been generated using AI.

Full size image

Theta band (4.0–7.5 Hz) topographic headmaps (left) and global mean amplitude (right) for Mindfulness Meditation and Paced Breathing conditions in each State (Pre-task [Pre], Task, Post-task [Post]). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Fig. 6

The alternative text for this image may have been generated using AI.

Full size image

Topographic headmaps (left) and global mean amplitude (right) depicting the differences between pre-task and task (Task − Pre) and pre-task and post-task (Post − Pre) for Mindfulness Meditation (MM) and Paced Breathing (PB) conditions in the theta band (4.0–7.5 Hz). Δ = change in global mean amplitude (µV). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Alpha appeared maximal in the occipital region in all states and both conditions (Fig. 7). Global alpha amplitude differed significantly between conditions when comparing task and pre-task (see Table 2 for supporting statistics). Specifically, there was a reduction in global alpha amplitude during the task in the mindfulness meditation condition, which appeared maximal at occipital sites as seen in Fig. 8. In contrast, the paced breathing condition showed an increase in global alpha amplitude during task, which appeared maximal at the vertex and left temporal region. There were no further significant changes in alpha between pre-task and post-task, or interactions.

Fig. 7

The alternative text for this image may have been generated using AI.

Full size image

Alpha band (8.0–13.0 Hz) topographic headmaps (left) and global mean amplitude (right) for Mindfulness Meditation and Paced Breathing conditions in each State (Pre-task [Pre], Task, Post-task [Post]). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Fig. 8

The alternative text for this image may have been generated using AI.

Full size image

Topographic headmaps (left) and global mean amplitude (right) depicting the differences between pre-task and task (Task − Pre) and pre-task and post-task (Post − Pre) for Mindfulness Meditation (MM) and Paced Breathing (PB) conditions in the alpha band (8.0–13.0 Hz). Δ = change in global mean amplitude (µV). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Beta appeared maximal at the vertex, parietal and occipital regions in all states and for both conditions (Fig. 9). There was a significant increase in global beta amplitude during the task relative to pre-task across conditions (see Table 2 for supporting statistics). As evidenced in Fig. 10, this also interacted significantly with condition, being somewhat reduced in mindfulness meditation, and with a large increase in paced breathing. No significant differences were identified between pre-task and post-task across conditions, and there was no interaction.

Fig. 9

The alternative text for this image may have been generated using AI.

Full size image

Beta band (13.5–30.0 Hz) topographic headmaps (left) and global mean amplitude (right) for Mindfulness Meditation and Paced Breathing conditions in each State (Pre-task [Pre], Task, Post-task [Post]). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Fig. 10

The alternative text for this image may have been generated using AI.

Full size image

Topographic headmaps (left) and global mean amplitude (right) depicting the differences between pre-task and task (Task − Pre) and pre-task and post-task (Post − Pre) for Mindfulness Meditation (MM) and Paced Breathing (PB) conditions in the beta band (13.5–30.0 Hz). Δ = change in global mean amplitude (µV). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Gamma appeared maximal at the occipital region but otherwise somewhat diffuse, and differed between conditions and states (Fig. 11). Global gamma amplitude during the task compared with pre-task differed significantly between conditions, somewhat decreased in mindfulness meditation, and increased in paced breathing as evidenced in Fig. 12 (see Table 2 for supporting statistics). No other significant differences were identified in gamma.

Fig. 11

The alternative text for this image may have been generated using AI.

Full size image

Gamma band (30.5–45.0 Hz) topographic headmaps (left) and global mean amplitude (right) for Mindfulness Meditation and Paced Breathing conditions in each State (Pre-task [Pre], Task, Post-task [Post]). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

Fig. 12

The alternative text for this image may have been generated using AI.

Full size image

Topographic headmaps (left) and global mean amplitude (right) depicting the differences between pre-task and task (Task − Pre) and pre-task and post-task (Post − Pre) for Mindfulness Meditation (MM) and Paced Breathing (PB) conditions in the gamma band (30.5–45.0 Hz). Δ = change in global mean amplitude (µV). Standard error of the mean is represented in the error bars. * indicates a significant effect (p < 0.05)

SCL

SCL was found to increase somewhat during the task compared with the pre-task, supported by statistics in Table 3. As shown in Fig. 13, this interacted significantly with condition, with a small decrease in mindfulness meditation, and an increase in paced breathing. There was also a significant increase in SCL post-task compared with pre-task, that did not interact with condition.

Table 3 Statistical results for skin conductance levels comparing Pre-task (Pre) vs. Task and Pre-task (Pre) vs. Post-task (Post), including interaction effects of Mindfulness Meditation (MM) vs. Paced Breathing (PB) conditions

Full size table

Fig. 13

The alternative text for this image may have been generated using AI.

Full size image

Mean SCL for Mindfulness Meditation (MM) and Paced Breathing (PB) conditions in each State (Pre-task [Pre], Task, and Post-task [Post]). Standard error of the mean is represented in the error bars. * p < 0.05

EEG Global Alpha and SCL Associations

Spearman’s rank-order correlation indicated no significant correlation between global alpha amplitude and SCL across conditions for pre-task, task, or post-task (all p ≥ 0.384). Further, the differences in global alpha amplitude and SCL between the pre-task and the task (task minus pre-task), as well as between the pre-task and the post-task (post-task minus pre-task), did not correlate (all p ≥ 0.147).

Discussion

This study investigated the effects of mindfulness meditation and paced breathing on neural oscillations and autonomic nervous system activity by measuring changes in global EEG amplitude and SCL. Respiratory rate decreased significantly during the task in both conditions, with a greater reduction observed in the paced breathing condition. Paced breathing increased global EEG amplitudes from pre-task to task across all frequency bands, as well as SCL. In contrast, mindfulness meditation reduced global alpha amplitude, while changes in the remaining EEG frequency bands and SCL were minimal. Additionally, no associations were found between global alpha amplitude and SCL in either condition. These results suggest that paced breathing broadly increases neural oscillations and sympathetic arousal, whereas mindfulness meditation reduces global alpha oscillations independently of sympathetic arousal, highlighting distinct psychophysiological differences between the two practices.

Firstly, the topographies across all EEG frequency bands in each state appeared relatively similar between the mindfulness meditation and paced breathing conditions, with the most discernible differences observed in the gamma oscillations. These topographies were similar to previous research that has investigated eyes-closed resting (Cave & Barry, 2021), mindfulness meditation (Cahn et al., 2010; Duda et al., 2024a, 2024b), and paced breathing (Park & Park, 2012) conditions. However, in relation to paced breathing, the specific EEG frequency bands and regions of interest analysed here represent a novel aspect of this study.

Both delta and theta oscillations showed a similar pattern of change, increasing during both mindfulness meditation and paced breathing, with the increase being more pronounced during paced breathing. The observed increase in delta oscillations may reflect heightened attentional engagement, supported by tentative findings in previous research (Lee et al., 2018; Lomas et al., 2015). Further, the increase in theta oscillations during mindfulness meditation and paced breathing aligns with past research linking theta activity to attention (Karakaş, 2020; Takahashi et al., 2005) and emotion regulation (Knyazev, 2007). Theta oscillations have been more frequently associated with meditation research, where increases are thought to reflect heightened awareness and focused attention (Aftanas & Golocheikine, 2001; Baijal & Srinivasan, 2010; Cahn et al., 2010). In support, Rodriguez-Larios et al. (2024) demonstrated that increases in theta activity with a central and midline distribution occur specifically during focused attention states, consistent with the observations in the present study. Together, these findings suggest that both practices involve similar overlapping attentional processes, which could account for the similarities observed in delta and theta oscillations. Discrepancies between our findings and those reported in Fumoto et al. (2004), Yu et al. (2011), and Park and Park (2012), which observed reductions in theta power during slow breathing, may be due to differences in experimental design and analyses performed.

Further, the increase in delta and theta oscillations observed post-task may reflect a prolonged relaxation effect following both mindfulness meditation and paced breathing. Elevated delta has been associated with deep relaxation but may also indicate fatigue or drowsiness (Fujiwara et al., 2022; Kiroy et al., 1996; Knyazev, 2012; Lal & Craig, 2002, 2005). While Cahn et al. (2010) found that self-reported drowsiness correlated with delta increases during meditation, the interpretation of delta activity remains nuanced. Dunn et al. (1999) reported greater delta activity during mindfulness meditation compared to a concentration state, yet delta was greatest during a relaxation state, suggesting that the delta increase in the present study may reflect relaxation rather than drowsiness. The post-task increase in theta oscillations may indicate a prolonged state of heightened attention, similar to the state effects observed during both mindfulness meditation and paced breathing, given the association between theta oscillations and attention (Klimesch et al., 2005; Sauseng et al., 2010). These findings suggest that delta activity in meditation may reflect a complex interplay between relaxation and attention, while increases in theta continue to support its established role in attentional processes common to both mindfulness and paced breathing.

Mindfulness meditation was associated with a reduction in alpha oscillations, consistent with previous findings (Amihai & Kozhevnikov, 2014; Duda et al., 2024a, 2024b, 2025; Rodriguez-Larios et al., 2021). This reduction has been linked to enhanced attentional focus and decreased mind-wandering in experienced meditators (Rodriguez-Larios et al., 2021). In contrast, paced breathing was associated with an increase in alpha oscillations, consistent with several studies (Fumoto et al., 2004; Park & Park, 2012; Stancák et al., 1993). Although both practices involve attentional engagement, the opposing effects on alpha oscillations suggest that mindfulness-specific dimensions, such as non-judgmental awareness and open monitoring, may engage distinct neural mechanisms. These mechanisms likely involve different aspects of inhibitory and attentional processes and contribute to the emotional and affective benefits associated with these practices (Lutz et al., 2008). Overall, these effects appear to be state-specific, emerging during mindfulness meditation and paced breathing but not persisting beyond it.

Beta oscillations increased significantly during paced breathing and decreased slightly during mindfulness meditation. Although previous studies have reported increases in beta oscillations during meditation (Ahani et al., 2014; Dunn et al., 1999; Kakumanu et al., 2018), and paced breathing (Stancák et al., 1993), findings across the literature remain mixed overall (Lomas et al., 2015; Park & Park, 2012). This variability may reflect differences in meditation types and experimental designs, as well as inconsistencies in the frequency ranges used to define beta activity, which can vary by up to 10 Hz across studies. Beta oscillations have been linked to attentional processes (Gola et al., 2012, 2013; Güntekin et al., 2013; Kamiński et al., 2012), which may account for the increases observed during paced breathing. However, beta oscillations do not appear to play a prominent role in mindfulness meditation and may be more relevant to concentration-based meditative practices, as suggested by their increase during paced breathing.

Further, similar to the pattern observed with beta oscillations, gamma oscillations significantly increased during paced breathing but showed a modest decrease during mindfulness meditation. Although there is currently no consensus on the functional role of gamma oscillations, they have been associated with increased attention, perception, and cognitive functioning, marking a heightened meditative state observed in experienced practitioners (Braboszcz et al., 2017; Cahn et al., 2010; Lee et al., 2018). In the present study, the mindfulness meditation was novel to participants, which supports the view that increases in gamma oscillations are less likely to occur in novice practitioners. In contrast, the increases observed during paced breathing may reflect its simplicity, as it does not require extensive practice to follow and perform consistently, unlike mindfulness meditation, which is more nuanced and can require extensive practice to achieve specific meditative states. Further, Amihai and Kozhevnikov (2014) reported mixed findings across different types of meditation, observing no change in gamma activity during Kasina meditation, increased gamma activity in the left hemisphere during Vipassana meditation, and decreased gamma activity during Deity and Rig-pa meditation, all compared to rest. These results suggest that increases in gamma activity may be specific to certain meditation practices. However, significant discrepancies across studies in the frequency ranges and topographic regions investigated make it difficult to draw definitive conclusions about the function of gamma oscillations. In the present study, the diverse topographic spread and variation of gamma oscillations across tasks underscore the need for further investigation, particularly given the technical challenges (e.g., artefact contamination) that often complicate the interpretation of gamma oscillations.

Alongside changes in neural oscillations, mindfulness meditation was associated with a modest reduction in SCL, whereas paced breathing led to an increase, suggesting that the two practices alter sympathetic arousal in distinct ways. While previous studies have linked mindfulness meditation to reductions in arousal, supporting its benefits for stress and emotion regulation (Hölzel et al., 2011; Scavone et al., 2020; Travis & Wallace, 1999), the relatively stable SCL observed here during mindfulness meditation suggests a state of 'relaxed alertness', neither under- or over-aroused (Britton et al., 2014). In contrast, the SCL increase observed during paced breathing diverges from our expectations and earlier findings that have reported reductions in arousal (Harris et al., 1976; Wielgosz et al., 2016). The rise in SCL during paced breathing may reflect increased attentional engagement rather than heightened arousal, as supported by the concurrent increase in alpha oscillations and evidence that SCL can index attention (Dawson et al., 2016). Additionally, Balban et al. (2023) reported differing effects on autonomic activity when comparing mindfulness meditation with various breathwork techniques, supporting the idea that SCL may index arousal in some contexts and attentional engagement in others. Together, these results support the idea that different practices engage distinct physiological mechanisms, potentially contributing to their unique benefits.

Further, sympathetic arousal increased post-task following both mindfulness meditation and paced breathing, potentially reflecting heightened alertness. This aligns with the observed rise in theta oscillations, which has been linked to attentional processes. Although increased physiological arousal is typically associated with heightened stress and reduced well-being, SCL has also been linked to emotion, cognition, and attention processes (Critchley, 2002). Thus, the observed increases in arousal may indicate a state of enhanced attention, highlighting the importance of interpreting SCL changes in context and underscoring the need for further investigation to clarify these effects.

Moreover, although SCL and alpha oscillations showed a similar directional change, they did not correlate in any state or condition, contrasting with the typically reported inverse relationship between these measures in resting states (Barry et al., 2005, 2020). The absence of this correlation suggests that the relationship between neural oscillations and sympathetic arousal may be more complex than previously hypothesised. One possible explanation is the study’s exclusive focus on eyes-closed states, unlike previous research that examined both eyes-open and eyes-closed states. Further, while the variance in alpha oscillations attributed to arousal in resting states is generally small, approximately 12%, this effect may be even weaker in studies limited to eyes closed states. Barry et al. (2007) also found that correlations between alpha oscillations and SCL diminish over time as the protocol progresses, although the reasons for this remain unclear. However, as no correlation was observed during either mindfulness meditation or paced breathing, the changes identified in the present study are unlikely to be driven by an arousal mechanism.

In summary, these findings indicate that mindfulness meditation and paced breathing have different effects on neural oscillations and sympathetic arousal, likely contributing to their respective health benefits. Specifically, paced breathing led to increases in global EEG amplitudes across all frequency bands and increased SCL, whereas mindfulness meditation resulted in a decrease in global alpha amplitude, with other frequency bands and SCL remaining relatively stable. Following both practices, there was an increase in delta and theta oscillations, accompanied by elevated SCL, indicating potential prolonged effects. Further research into the connections between neural oscillations and autonomic regulation in mindfulness meditation and breathing exercises could deepen our understanding of their impact on cognitive and physiological processes. Such knowledge will be vital for developing precise, evidence-based interventions to optimise mental and physical health, given that these practices modulate attention and arousal through distinct neural and autonomic pathways.

Limitations & Future Directions

This study’s focus on a healthy young adult population limits the generalisability of findings, particularly regarding clinical applications for psychological disorders. Newson and Thiagarajan (2019) highlighted that EEG activity varies across all frequency bands in different psychiatric disorders, suggesting certain mindfulness interventions may be more beneficial for specific conditions. Future research should investigate these practices in clinical populations, ideally through longitudinal or intervention-based designs that can clarify their long-term effects on brain function and autonomic regulation. Further, variability in physiological responses, as well as discrepancies with previous findings, may be influenced by factors such as stress, drowsiness, boredom, or difficulty engaging with the task. Incorporating behavioural and psychological measures will also be essential for capturing the full scope of mindfulness-based benefits and informing more tailored interventions.

Global EEG measures were utilised in the present study based on the rationale that global alpha activity serves as an inverse marker of arousal. However, future research may benefit from incorporating topographic analyses to gain more detailed insights into localised neural mechanisms underlying meditation-related cognitive processes and state changes that global measures may not be fully capture. This approach is particularly important given the topographic differences between mindfulness meditation and paced breathing observed in the present study, as well as the varying topographic effects reported in previous meditation research (Cahn & Polich, 2006; Lee et al., 2018; Lomas et al., 2015; Tang et al., 2015).

Furthermore, as evident in the present study, while some studies have reported changes in SCL, specific effects are not always observed, suggesting a need to explore alternative indices of autonomic activity, such as heart rate or heart rate variability (HRV). Unlike SCL, HRV provides valuable insights into both the sympathetic and parasympathetic branches of the autonomic nervous system and is directly influenced by respiration rate, which may be especially relevant in mindfulness meditation and paced breathing practices (Shaffer & Ginsberg, 2017). Examining both SCL and HRV can thus enhance our understanding of the distinct autonomic mechanisms engaged during mindfulness meditation and paced breathing, as these measures capture differing physiological effects (Dawson et al., 2016). To support this, future research should also incorporate a direct measure of respiration rate to ensure task adherence and better understand its relationship in shaping neural and autonomic changes.

Future studies should combine paced breathing with mindfulness meditation to control for breathing rate, as well as separately manipulate breathing rate to assess its impact on alpha oscillations. These approaches would help determine whether breathing, attentional focus, or their interaction drives alpha changes, clarifying mechanisms underlying both practices. Lastly, not only does attention to the breath matter, but the type of breathing, particularly nasal versus mouth breathing, which was not controlled for in the present study. Differences in brain activity modulation and autonomic system changes have been associated with nasal versus mouth breathing, which may partly explain the differences observed in this study (Heck et al., 2017; Zaccaro et al., 2022; Zelano et al., 2016).

Data Availability

The datasets generated and/or analysed during this study can be obtained from the corresponding author on reasonable request.

References

• Aftanas, L. I., & Golocheikine, S. A. (2001). Human anterior and frontal midline theta and lower alpha reflect emotionally positive state and internalized attention: High-resolution EEG investigation of meditation. Neuroscience Letters, 310(1), 57–60. https://doi.org/10.1016/S0304-3940(01)02094-8

  Article PubMed Google Scholar 

• Ahani, A., Wahbeh, H., Nezamfar, H., Miller, M., Erdogmus, D., & Oken, B. (2014). Quantitative change of EEG and respiration signals during mindfulness meditation. Journal of Neuroengineering and Rehabilitation, 11, 1–11. https://doi.org/10.1186/1743-0003-11-87