Stress is visible in the brain during video meetings: in a Microsoft Human Factors Lab EEG study, beta-wave activity linked to stress kept building across four back-to-back 30-minute calls, and 10-minute breaks let it drop and reset (Microsoft, Work Trend Index brain research 2021). A 2023 laboratory experiment reached a similar verdict with a different method: 50 minutes of videoconferencing raised frontal theta and alpha power and lowered heart rate compared with the identical face-to-face session, signals the authors interpret as fatigue (Riedl et al., Scientific Reports 2023). Self-report data point the same way: 13.8% of women and 5.5% of men told Stanford researchers they felt very to extremely fatigued in a survey of 10,322 people. Meeting load is the multiplier, which our meeting overload statistics track from the calendar side. We aggregated data from the Microsoft Human Factors Lab, the Stanford Virtual Human Interaction Lab, peer-reviewed EEG and ECG experiments in Scientific Reports and Sensors, field experiments in the Journal of Applied Psychology, the Microsoft Work Trend Index and the other primary sources listed in the methodology.
TL;DR
- Beta-wave stress activity accumulated across four back-to-back 30-minute meetings; 10-minute breaks reset it (Microsoft Human Factors Lab, 2021).
- 50 minutes of videoconferencing raised frontal theta and alpha power vs the same lecture face to face (Riedl et al., Scientific Reports 2023).
- Heart rate declined steadily during a 50-minute video session (slope -0.1482, p = 0.0003) (Riedl et al., Scientific Reports 2023).
- Pupil dilation was significantly larger in in-person than video conversations (Tremmel et al., Sensors 2025).
- Each extra hour of videoconferencing cost heavy users 17 minutes of sleep (Onyper, Nature and Science of Sleep 2024).
- 13.8% of women vs 5.5% of men felt very to extremely Zoom-fatigued (Stanford, survey of 10,322, 2021).
- Video meeting fatigue was 14.9% higher for women and 11.1% higher for Asian than White participants (Ratan, Miller and Bailenson, 2022).
- Turning self-view off cut student fatigue from 3.41 to 2.25 (d = -0.71) (Basch, Albus and Seufert, Scientific Reports 2025).
- Camera use raised daily fatigue in a 4-week field experiment with 103 employees (Shockley et al., Journal of Applied Psychology 2021).
- Videoconference fatigue correlated r = .588 with depression and r = .511 with burnout (Montag et al., 2022).
- Weekly Teams meeting time rose 252% from February 2020 to February 2022 (Microsoft Work Trend Index 2022).
- 57% of meetings are ad hoc calls without a calendar invite (Microsoft Work Trend Index 2025).
1. Brain-Wave Evidence: What EEG Shows During Video Calls
EEG moved the Zoom fatigue debate from opinion to measurement. In Microsoft’s Human Factors Lab study, 14 participants wore EEG caps through four 30-minute meetings (two continuous hours); without breaks, beta-wave activity associated with stress accumulated, while 10-minute breaks between meetings let it drop (Microsoft WorkLab, Research Proves Your Brain Needs Breaks). The breaks also produced positive frontal alpha asymmetry, a marker the lab tied to engagement, versus negative asymmetry (withdrawal) in the back-to-back condition, and the transitions between meetings themselves caused stress spikes.
The strongest controlled evidence comes from Austria. In a within-subjects lab experiment, 35 students (20 men, 15 women, mean age 24.06) attended the same 50-minute lecture once by videoconference and once in person, a week apart. Videoconferencing produced larger frontal theta power that rose steadily through the lecture, more pronounced alpha power at Fz, Pz, O1 and O2, and a significantly higher frontal theta/beta ratio (Riedl, Kostoglou, Wriessnegger and Müller-Putz, Scientific Reports 2023). The point is not that one band moved; it is that several independent markers of drowsiness and reduced alertness moved together in the video condition only.
| Metric | Value | Source |
|---|---|---|
| EEG participants, back-to-back meeting study | 14, four 30-minute meetings per session | Microsoft Human Factors Lab, March 2021 |
| Break length tested | 10 minutes between meetings | Microsoft Human Factors Lab, 2021 |
| Beta-wave (stress) activity | Accumulated without breaks; dropped with breaks | Microsoft Human Factors Lab, 2021 |
| Lab experiment sample | 35 (20 male, 15 female), mean age 24.06 | Riedl et al., Scientific Reports 2023 |
| Session length compared | 50-minute lecture, video vs face to face | Riedl et al., Scientific Reports 2023 |
| Frontal theta (Fz) during video | Larger, rising steadily (slope p = 0.0145) | Riedl et al., Scientific Reports 2023 |
| Theta/beta ratio, frontal | Significantly greater during video (p < 0.05) | Riedl et al., Scientific Reports 2023 |
| Self-view and alpha activity | Significantly greater alpha with self-view on (32 participants, 20-minute session) | Xu et al., Cyberpsychology, Behavior, and Social Networking 2024 |
Context note: the Microsoft study is small (14 participants) and was published as an industry report, not a peer-reviewed paper. The Irish EEG study by Xu, Whelan, O’Brien and O’Hora (16 men, 16 women) found that alpha activity did not differ by gender or over the 20-minute session (Cyberpsychology, Behavior, and Social Networking 2024).
2. Heart, Eyes and Sleep: The Body Signals Beyond the Brain
Cardiac data in the same Austrian experiment told a story of declining arousal. Heart rate fell gradually throughout the videoconference session (slope -0.1482, p = 0.0003), and participants in the face-to-face condition showed significantly higher heart rate, LFnu and ln[LF/HF] (Riedl et al., Scientific Reports 2023). The video session instead showed higher pNN50, RMSSD and SDRR, heart-rate-variability measures that typically rise as the body moves toward a parasympathetic, low-activation state. Paired with the EEG, the authors conclude the video format produced fatigue, not relaxation.
A 2025 multimodal study adds the eyes. Recording eye tracking, head and hand movement, heart rate, respiration and EEG in 18 adults, researchers at the University of Southampton found that pupils dilated significantly more in person, and horizontal gaze variance and lateral head motion were significantly reduced on video, the price of staying inside the camera frame (Tremmel et al., Sensors 2025). The fatigue also follows people to bed: in an 8-day diary study of 96 young adults, heavier video-call users lost 17 minutes of sleep and nearly 1% of sleep efficiency for every additional hour of calls (Onyper, Nature and Science of Sleep 2024).
| Metric | Value | Source |
|---|---|---|
| Heart rate trend during 50-minute video session | Gradual decline, slope -0.1482 (p = 0.0003) | Riedl et al., Scientific Reports 2023 |
| HRV during video vs face to face | Significantly higher pNN50, RMSSD and SDRR | Riedl et al., Scientific Reports 2023 |
| Pupil dilation | Significantly larger in person (p < 0.05) | Tremmel et al., Sensors 2025 |
| Multimodal study sample | 18 adults, 5-minute reading and free-talk tasks per condition | Tremmel et al., Sensors 2025 |
| Heart rate, in person vs video | Higher in person, not statistically significant | Tremmel et al., Sensors 2025 |
| Sleep lost per extra hour of video calls (heavy users) | 17 minutes; sleep efficiency b = -0.88 (nearly 1%) | Onyper, Nature and Science of Sleep 2024 |
| Poor sleepers (PSQI above 5) | 71% of high-volume vs 46% of low-volume callers | Onyper, Nature and Science of Sleep 2024 |
| Average daily video-call load, students | 1.3 hours (high group 2.0, low group 0.7) | Onyper, Nature and Science of Sleep 2024 |
Context note: biomarker evidence of chronic stress is weaker than evidence of acute fatigue. In a 2025 study of 102 remote “smart workers”, perceived technostress creators showed little association with hair cortisol, the long-term stress biomarker, except through an interaction with supervisor support (Girardi et al., PLoS One 2025).
3. Measuring the Fatigue: Scales, Prevalence and Who Is Hit Hardest
Before biomarkers, the field needed a ruler. Stanford researchers built the Zoom Exhaustion & Fatigue (ZEF) Scale by cutting 49 candidate items to 15 across five dimensions (general, social, emotional, visual and motivational fatigue), then validated it on 2,724 respondents (Fauville et al., Computers in Human Behavior Reports 2021). In that validation, more frequent, longer and more “bursty” meetings were each associated with higher fatigue. The scale is now the common currency of the field and was used inside the Riedl EEG experiment, where ZEF scores were significantly higher after the video lecture.
The first large survey with the scale found a sharp gender gap. Of 10,322 participants, 13.8% of women vs 5.5% of men reported feeling very to extremely fatigued, roughly 1 in 7 women vs 1 in 20 men (Stanford News, April 2021). Women had the same number of meetings as men, but their meetings ran longer and they took fewer breaks between them. The peer-reviewed version, on 9,787 participants, attributes the gap to five nonverbal mechanisms: mirror anxiety, hyper-gaze (feeling watched), physical confinement, the effort of producing nonverbal cues and the effort of monitoring others’ cues (Fauville et al., Computers in Human Behavior Reports 2023).
| Metric | Value | Source |
|---|---|---|
| ZEF Scale items | 49 candidate items reduced to 15 | Fauville et al., Computers in Human Behavior Reports 2021 |
| ZEF development and validation samples | 395 and 2,724 respondents | Fauville et al., 2021 |
| Very to extremely fatigued, women vs men | 13.8% vs 5.5% (n = 10,322) | Stanford Virtual Human Interaction Lab, April 2021 |
| Peer-reviewed sample on gender mechanisms | 9,787 participants | Fauville et al., Computers in Human Behavior Reports 2023 |
| Fatigue gap, women vs men | 14.9% higher for women (2.77 vs 2.41) | Ratan, Miller and Bailenson, Cyberpsychology, Behavior, and Social Networking 2022 |
| Fatigue gap, Asian vs White participants | 11.1% higher (2.77 vs 2.50) | Ratan, Miller and Bailenson, 2022 |
| Facial dissatisfaction vs fatigue | r = 0.26 (n = 613) | Ratan, Miller and Bailenson, 2022 |
| ZEF score, women vs men (German-speaking sample) | 38.27 vs 33.58, d = 0.41 (n = 311) | Montag et al., Journal of Affective Disorders Reports 2022 |
Context note: Ratan and colleagues found the gender and race differences were mediated by dissatisfaction with one’s own facial appearance (PMC full text). The Stanford survey also reported higher fatigue among younger people, introverts and less emotionally stable respondents. Most recent available data on large-sample prevalence: Stanford, 2021 to 2023.
4. Camera, Self-View and Backgrounds: What Actually Moves the Needle
The theory behind these experiments comes from Jeremy Bailenson’s “nonverbal overload” paper, which proposed four interface-driven causes of Zoom fatigue and has been cited more than 1,200 times according to OpenAlex (Bailenson, Technology, Mind, and Behavior 2021). The cleanest experimental lever turned out to be the mirror. When students switched self-view off, fatigue fell from 3.41 to 2.25 and extraneous cognitive load from 2.52 to 1.64, effect sizes of d = -0.71 and d = -0.82 (Basch, Albus and Seufert, Scientific Reports 2025). Active participation also lowered fatigue (3.06 vs 3.58), while virtual backgrounds raised cognitive load without changing fatigue.
Camera-on rules are less settled than the headlines suggest. A 4-week within-person field experiment that manipulated camera use found it linked to higher daily fatigue, which in turn related to lower voice and engagement in meetings; the authors report support for a model in which cameras are more fatiguing for women and newer employees (Shockley et al., Journal of Applied Psychology 2021). But two preregistered studies found camera use unrelated to Zoom fatigue, while fatigue itself predicted more conformity to majority opinions (Masjutin et al., Scientific Reports 2024). The practical reading: self-view is the better-supported target than the camera itself.
| Metric | Value | Source |
|---|---|---|
| Fatigue, self-view off vs on | 2.25 vs 3.41 (d = -0.71, n = 25) | Basch, Albus and Seufert, Scientific Reports 2025 |
| Extraneous cognitive load, self-view off vs on | 1.64 vs 2.52 (d = -0.82) | Basch, Albus and Seufert, 2025 |
| Fatigue, active participation vs none | 3.06 vs 3.58 (d = -0.39, n = 28) | Basch, Albus and Seufert, 2025 |
| Extraneous load, virtual vs natural background | 3.14 vs 2.45 (d = 0.59); fatigue not significant | Basch, Albus and Seufert, 2025 |
| Camera field experiment | 103 employees, 1,408 daily observations, 4 weeks | Shockley et al., Journal of Applied Psychology 2021 |
| Camera use and fatigue, preregistered studies | No relationship (n = 287 and n = 64) | Masjutin et al., Scientific Reports 2024 |
| Virtual background survey | 610 users; background type and content linked to fatigue | Li and Zhang, Frontiers in Psychology 2024 |
Context note: the Basch study followed 39 psychology students at a German university across a semester, with smaller subsamples per intervention (n = 12 to 28), so its effect sizes are promising but not definitive. Li and Zhang’s background findings are cross-sectional (Frontiers in Psychology 2024).
5. Meeting Load: The Exposure Behind the Biomarkers
Lab studies measure one session; the workweek stacks them. Weekly meeting time for the average Microsoft Teams user rose 252% between February 2020 and February 2022, and the number of weekly meetings rose 153% (Microsoft Work Trend Index 2022). A year earlier, Microsoft had measured time in Teams meetings at 2.5 times its pre-pandemic level and the average meeting lengthening from 35 to 45 minutes (Work Trend Index 2021). That longer meeting is the key link to the lab data: the Riedl experiment found fatigue markers after 50 minutes.
The load has not normalized. In 2023 Microsoft said meetings and calls per week had roughly tripled since February 2020 (192% growth), that 23% of work time went to Teams meetings and that 68% of people lacked enough uninterrupted focus time (Work Trend Index 2023). The 2025 telemetry adds fragmentation: 57% of meetings are ad hoc calls without an invite and meetings after 8 pm rose 16% year over year (Work Trend Index 2025). Unscheduled calls make planned breaks harder to protect, and breaks are what reset beta-wave stress in the lab. The dollar side of this load is covered in our meeting cost statistics.
| Metric | Value | Source |
|---|---|---|
| Weekly Teams meeting time, Feb 2020 to Feb 2022 | +252% (weekly meetings +153%) | Microsoft Work Trend Index 2022 |
| Time in Teams meetings, Feb 2020 to Feb 2021 | 2.5x | Microsoft Work Trend Index 2021 |
| Average meeting length | 35 to 45 minutes | Microsoft Work Trend Index 2021 |
| Meetings under 15 minutes | 60% of meetings (Feb 2022) | Microsoft Work Trend Index 2022 |
| Share of work time in Teams meetings | 23% | Microsoft Work Trend Index 2023 |
| Workers lacking uninterrupted focus time | 68% (31,000 surveyed) | Microsoft Work Trend Index 2023 |
| Ad hoc meetings without a calendar invite | 57% | Microsoft Work Trend Index 2025 |
| Meetings after 8 pm, year over year | +16% | Microsoft Work Trend Index 2025 |
Context note: Microsoft’s meeting figures are Microsoft 365 telemetry from Teams users, not a representative sample of all workers. The 2025 report also counts interruptions every 2 minutes (275 a day). Tools that shorten meetings or replace them with summaries are tracked in our AI meeting assistant statistics.
6. Fatigue, Burnout and Mental Health Links
The biomarkers matter because fatigue does not stay inside the meeting. In a sample of 311 adults, videoconference fatigue correlated r = .588 with depression and r = .511 with burnout, both at p < .001 (Montag, Rozgonjuk, Riedl and Sindermann, Journal of Affective Disorders Reports 2022). The authors’ mediation models suggest videoconference fatigue may partly carry the link between neuroticism and both outcomes, though the cross-sectional design cannot settle direction. Wider workforce data show the same tension: 54% of workers felt overworked and 39% exhausted in Microsoft’s 2021 survey of 31,092 people.
Student data show both the scale and the persistence of the problem. Among 386 Thai medical students, 9.6% had high Zoom fatigue, and more online sessions predicted it while regular exercise was protective (Charoenporn et al., F1000Research 2024). In Turkey, a two-wave study of 351 students now back in face-to-face classes found Zoom fatigue still lowered subjective vitality six months apart (Buyukcebeci, Frontiers in Psychology 2025). For the broader workforce picture, see our workplace burnout statistics.
| Metric | Value | Source |
|---|---|---|
| VC fatigue vs depression | r = .588 (p < .001, n = 311) | Montag et al., Journal of Affective Disorders Reports 2022 |
| VC fatigue vs burnout | r = .511 (p < .001) | Montag et al., 2022 |
| VC fatigue vs neuroticism | r = .395 (p < .001) | Montag et al., 2022 |
| High Zoom fatigue prevalence, medical students | 9.6% (n = 386) | Charoenporn et al., F1000Research 2024 |
| Depressive disorder in same sample | 61.9%, linked to Zoom fatigue (p = 0.004) | Charoenporn et al., 2024 |
| Faculty fatigue score, Philippines | 3.35 of 5 (n = 322) | Oducado et al., Education and Information Technologies 2022 |
| Predictors, faculty | Duration beta 0.166; interval between meetings beta -0.085; model explains 30.7% | Oducado et al., 2022 |
| Workers feeling overworked / exhausted | 54% / 39% (n = 31,092) | Microsoft Work Trend Index 2021 |
Context note: the Oducado faculty study (Education and Information Technologies 2022) independently confirms the break effect seen in EEG: longer intervals between videoconferences predicted lower fatigue, and longer sessions predicted higher fatigue. Prevalence figures are not comparable across studies because cut-offs on the ZEF scale differ.
Summary: Video Meeting Fatigue by the Numbers
| Metric | Value | Source |
|---|---|---|
| Beta-wave stress, back-to-back meetings | Accumulated; reset with 10-minute breaks | Microsoft Human Factors Lab, 2021 |
| EEG fatigue markers after 50 minutes of video | Higher frontal theta, alpha and theta/beta ratio | Riedl et al., Scientific Reports 2023 |
| Heart rate during 50-minute video session | Steady decline (slope -0.1482) | Riedl et al., Scientific Reports 2023 |
| Pupil dilation, in person vs video | Significantly larger in person | Tremmel et al., Sensors 2025 |
| Sleep lost per extra hour of calls (heavy users) | 17 minutes | Onyper, Nature and Science of Sleep 2024 |
| Poor sleepers, high vs low call volume | 71% vs 46% | Onyper, 2024 |
| Very to extremely fatigued, women vs men | 13.8% vs 5.5% | Stanford, 2021 (n = 10,322) |
| Fatigue gap, women vs men | 14.9% higher | Ratan, Miller and Bailenson, 2022 |
| ZEF Scale | 15 items, 5 dimensions, validated on 2,724 | Fauville et al., 2021 |
| Gender mechanisms sample | 9,787 participants, 5 nonverbal mechanisms | Fauville et al., 2023 |
| Fatigue, self-view off vs on | 2.25 vs 3.41 (d = -0.71) | Basch, Albus and Seufert, 2025 |
| Camera field experiment | 103 employees, 1,408 daily observations | Shockley et al., 2021 |
| Camera use and fatigue, preregistered | No relationship (n = 287, n = 64) | Masjutin et al., 2024 |
| Weekly Teams meeting time, 2020 to 2022 | +252% | Microsoft Work Trend Index 2022 |
| Average meeting length | 35 to 45 minutes | Microsoft Work Trend Index 2021 |
| Ad hoc meetings | 57% | Microsoft Work Trend Index 2025 |
| VC fatigue vs depression | r = .588 | Montag et al., 2022 |
| VC fatigue vs burnout | r = .511 | Montag et al., 2022 |
| High Zoom fatigue, medical students | 9.6% | Charoenporn et al., 2024 |
Methodology and Sources
Every figure above was read during research for this article in a peer-reviewed paper (abstract or open full text via PubMed Central and Europe PMC), a university news release or a Microsoft research report. Physiological findings come from small laboratory samples and are reported with their sample sizes; statistics circulating on blogs without a primary document were excluded.
- Microsoft WorkLab: Research Proves Your Brain Needs Breaks (Human Factors Lab EEG study, 2021)
- Riedl R, Kostoglou K, Wriessnegger SC, Müller-Putz GR: Videoconference fatigue from a neurophysiological perspective (EEG and ECG), Scientific Reports 2023
- Xu J, Whelan E, O’Brien A, O’Hora D: Does Self-View Mode Generate More Videoconferencing Fatigue in Women than Men? An Experiment Using EEG Signals, Cyberpsychology, Behavior, and Social Networking 2024
- Tremmel C et al.: Physiological and Behavioral Response Differences Between Video-Mediated and In-Person Interaction, Sensors 2025
- Onyper S: Videoconferencing During the COVID-19 Pandemic is Associated with Sleep Disruption in Young Adults, Nature and Science of Sleep 2024
- Girardi D et al.: Technostress creators, hair cortisol and strain in remote work, PLoS One 2025
- Stanford Virtual Human Interaction Lab: Stanford News, Zoom fatigue worse for women (April 2021); Fauville et al., Zoom Exhaustion & Fatigue Scale, Computers in Human Behavior Reports 2021; Fauville et al., Video-conferencing usage dynamics and nonverbal mechanisms exacerbate Zoom Fatigue, particularly for women, Computers in Human Behavior Reports 2023; Bailenson JN, Nonverbal overload, Technology, Mind, and Behavior 2021
- Ratan R, Miller DB, Bailenson JN: Facial Appearance Dissatisfaction Explains Differences in Zoom Fatigue, Cyberpsychology, Behavior, and Social Networking 2022
- Basch JM, Albus P, Seufert T: Fighting Zoom fatigue: Evidence-based approaches in university online education, Scientific Reports 2025
- Shockley KM et al.: The fatiguing effects of camera use in virtual meetings, Journal of Applied Psychology 2021
- Masjutin L et al.: Fatigued individuals show increased conformity in virtual meetings, Scientific Reports 2024
- Li BJ, Zhang H: Virtual background use and videoconference fatigue, Frontiers in Psychology 2024
- Montag C et al.: Videoconference fatigue, burnout and depression, Journal of Affective Disorders Reports 2022
- Charoenporn V et al.: Zoom fatigue among medical students in Thailand, F1000Research 2024; Buyukcebeci A: Post-pandemic Zoom fatigue and subjective vitality, Frontiers in Psychology 2025; Oducado RMF et al.: Factors predicting videoconferencing fatigue among higher education faculty, 2022
- Microsoft Work Trend Index: 2021, The Next Great Disruption Is Hybrid Work; 2022, Great Expectations; 2023, Will AI Fix Work?; 2025, Breaking Down the Infinite Workday
- Data watch: Physiological studies are small (14 to 35 participants for EEG, 18 for the multimodal study) and mostly use students, so they show that fatigue is measurable, not how common it is. The Microsoft EEG study was published as an industry report rather than in a peer-reviewed journal. The Stanford gender figures (13.8% vs 5.5%) come from the April 2021 Stanford News release on 10,322 respondents; the later peer-reviewed paper reports 9,787 participants, and its full text could not be read (ScienceDirect blocked access), so only its abstract-level findings are used. Bailenson’s 2021 paper and the Stanford ZEF paper were read through abstracts and metadata (the APA journal page returned 403); the citation count (1,208) is from OpenAlex. Shockley et al. and Masjutin et al. disagree on whether camera use itself causes fatigue; self-view results (Basch 2025, Xu 2024) are more consistent. Xu et al. found no gender difference in EEG alpha, in contrast with the survey-based gender gap. Microsoft meeting telemetry covers Teams users only and changes definitions between reports (2.5x in 2021, 252% in 2022, roughly 3x or 192% in 2023), so the series should not be chained. Most recent available data for large-sample prevalence: Stanford, 2021 to 2023; no newer national prevalence survey using the ZEF scale was found.
Last updated: October 3, 2026. We update this roundup quarterly, and the next refresh is expected when Microsoft publishes its 2026 Work Trend Index telemetry and new EEG or heart-rate-variability studies on videoconferencing appear in PubMed Central.