1. Neuroscience

Respiratory and cardiac interoceptive sensitivity in the first two years of life

  1. Markus R Tünte  Is a corresponding author
  2. Stefanie Hoehl
  3. Moritz Wunderwald
  4. Johannes Bullinger
  5. Asena Boyadziheva
  6. Lara Maister
  7. Birgit Elsner
  8. Manos Tsakiris
  9. Ezgi Kayhan
  1. University of Vienna, Faculty of Psychology, Department of Developmental and Educational Psychology, Austria
  2. Vienna Doctoral School Cognition, Behavior and Neuroscience, University of Vienna, Austria
  3. Ludwig-Maximilians-Universität München, Germany
  4. School of Human and Behavioural Sciences, College of Human Sciences, Prifysgol Bangor University, United Kingdom
  5. Department of Developmental Psychology, University of Potsdam, Germany
  6. Department of Psychology, Royal Holloway University of London, United Kingdom
https://doi.org/10.7554/eLife.91579.4
Share this article
Cite this article
  1. Markus R Tünte
  2. Stefanie Hoehl
  3. Moritz Wunderwald
  4. Johannes Bullinger
  5. Asena Boyadziheva
  6. Lara Maister
  7. Birgit Elsner
  8. Manos Tsakiris
  9. Ezgi Kayhan
(2025)
Respiratory and cardiac interoceptive sensitivity in the first two years of life
eLife 12:RP91579.
https://doi.org/10.7554/eLife.91579.4
  2. Download BibTeX
  3. Download .RIS
12 figures, 26 tables and 1 additional file

Figures

Figure 1
Download asset Open asset
Looking times for (A) iBEATs and (B) iBREATH in 9-mo infants.

Looking times for the (A) iBEATs (paired t-test, N = 52, t=–2.96, p=0.005) and (B) iBREATH (paired t-test, N = 56, t=–2.80, p=0.007) tasks. In both tasks, 9-mo-old infants looked significantly longer at stimuli presented synchronously to their own physiological signals. Black dots refer to the group mean. Black bars refer to the standard error of the mean. Gray lines and colorful dots refer to individual mean-looking times per condition and infant. In (A) blue represents the synchronous condition and orange is the asynchronous, while in (B) green represents the synchronous condition, and purple is the asynchronous.

Figure 2
Download asset Open asset
Results from the MEGA analysis for (A) iBEATs and (B) iBREATH.

Results from the MEGA analysis for (A) iBEATs (combined sample N = 135) and (B) iBREATH (combined sample N = 120). Plot of difference scores computed as mean synchronous looking times minus mean asynchronous looking times per individual for each age group, as well as the combined sample. Red dots refer to mean effects for the respective analysis as described above, red bars refer to 95% confidence/credible intervals. Dashed line indicates a difference of 0. For 3, 9, and 18 mo age groups our preregistered analysis is plotted. For the combined sample we computed a linear mixed model using lme4 for visualization purposes as results from a mixed model with a beta error distribution cannot easily be transformed back to the original scale.

Figure 3
Download asset Open asset
Relationship between iBEATs and iBREATH using a combined sample.

Histogram with a plotted line for individual performance on iBEATs and iBREATH using a beta regressionN = 84. Following Maister et al., 2017, individual difference scores were computed as the proportion of the absolute difference between synchronous and asynchronous trials.

Figure 4
Download asset Open asset
Exploratory analysis for age effect.

Absolute proportional scores for (A) iBEATs and (B) iBREATH plotted for each age group. Red dots refer to group means, and colorful dots to individual means, ** refers to a significant result (p<0.01).

Figure 5
Download asset Open asset
Specification curve analysis for the (A) iBEATs and (B) iBREATH task.

Specification curve analysis plotting standardized beta regression coefficients (y-axis) and number of analysis (x-axis) for (A) iBEATs and (B) iBREATH. Number of analysis (x-axis) are ordered increasing from lowest to highest standardized beta regression coefficient. Blue color indicates a significant effect (p<0.05) for a mean synchronous preference, red color indicates a significant effect (p<0.05) for a mean asynchronous preference, and gray indicates a non-significant outcome.

Appendix 1—figure 1
Download asset Open asset
Stimuli used for (A) iBEATs and (B) iBREATH.
Appendix 1—figure 2
Download asset Open asset
Mean differences in (A) iBEATs (blue refers to the synchronous and orange to the asynchronous condition) and (B) iBREATH (green refers to the synch and purple to the asynchronous condition) for the 9-mo-olds.

Gray dots and lines refer to individual infants, black dots and error bars to the results of mean comparison and corresponding 95% confidence intervals. Infants with alternative randomization are highlighted in red.

Appendix 1—figure 3
Download asset Open asset
Infants’ preference over time in the iBEATs paradigm for all age groups.

Here results of paired t-tests for the iBEATs with different inclusion criteria for trials are displayed for all age groups. The left column shows a cumulative analysis in which all previous trials are included. The right column shows a 3-trial sliding window analysis in which the three preceding and three proceeding trials are included. Blue indicates a significant result of the paired t-test.

Appendix 1—figure 4
Download asset Open asset
Infants’ preference over time in the iBREATH paradigm for all age groups.

Here results of paired t-tests for the iBREATH with different inclusion criteria for trials are displayed for all age groups. The left column shows a cumulative analysis in which all previous trials are included. The right column shows a 3-trial sliding window analysis in which the three preceding and three proceeding trials are included. Blue indicates a significant result of the paired t-test.

Appendix 2—figure 1
Download asset Open asset
Simulating data frames for sample sizes from 15 to 125 building up on the iBEATs data from the 9-mo-olds.

Results from the simulations. In (A) mean effects and 95% confidence intervals are plotted for the different sample sizes. Red color indicates a significant effect, while blue indicates a non-significant result. In (B) the percent of significant results are plotted with a fitted line.

Appendix 4—figure 1
Download asset Open asset
Descriptive results from the specification curve analysis for the iBEATs task.

Blue coloring in (A) and (B) refers to a significant result for a mean synchronous preference, while red color indicates to a significant result for a mean asynchronous preference (p<0.05) for the specification and test. In (A) standardized beta regression estimates are plotted. In (B) an overview for a range of analytical choices is given. In (C) analytical choices are further decomposed.

Appendix 4—figure 2
Download asset Open asset
Descriptive results from the specification curve analysis for the iBREATH task.

Blue coloring in (A) and (B) refers to a significant result for a mean synchronous preference, while red color indicates to a significant result for a mean asynchronous preference (p<0.05) for the specification and test. In (A) standardized beta regression estimates are plotted. In (B) an overview for a range of analytical choices is given. In (C) analytical choices are further decomposed.

Tables

Table 1
Interactions between condition and age for the iBEATs MEGA-analysis.
TermEstimateSEz-valuep-value
Condition * age (3- vs 9 mo)0.000.110.020.982
Condition * age (3- vs 18 mo)0.140.131.080.283
Condition * age (9- vs 18 mo)0.140.121.120.264

  1. Results for the mixed model using a beta error distribution. Here only the interaction between condition and age are reported for all combinations. Detailed results can be found in Appendix 3.

Table 2
Interactions between condition and age for the iBREATH MEGA-analysis.
TermEstimateSEz-valuep-value
Condition * age (3- vs 9 mo)0.020.120.170.864
Condition * age (3- vs 18 mo)0.230.161.500.134
Condition * age (9- vs 18 mo)0.210.151.430.154

  1. Results for the mixed model using a beta error distribution. Here only the interaction between condition and age are reported for all combinations. Detailed results can be found in Appendix 3.

Table 3
Effects of iBEATs on iBREATH for all age groups, as well as for interactions between iBEATs and age.
TermEstimateSEz-valuep-value
iBEATs (3 mo)–1.830.97–1.890.059
iBEATs (9 mo)–1.160.90–1.300.192
iBEATs (18 mo)1.301.021.270.204
iBEATs * age (3- vs 9 mo)–0.150.34–0.420.674
iBEATs * age (3- vs 18 mo)3.131.412.220.027
iBEATs * age (9- vs 18 mo)2.451.361.810.070

  1. Results for the beta regression of iBEATs scores on iBREATH scores. Detailed results can be found in Appendix 3.

Table 4
Change in absolute proportional scores across age groups for the iBREATH.
TermEstimateSEz-valuep-value
Intercept–1.610.1510.55<0001
9 mo–0.040.19–0.180.853
18 mo0.610.212.860.004

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 3 mo age group.

Table 5
Change in absolute proportional scores across age groups for the iBEATs.
TermEstimateSEz-valuep-value
Intercept–1.110.14–7.91<0.001
9 mo–0.040.19–0.220.826
18 mo–0.090.23–0.410.684

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 3 mo age group.

Table 6
Number of significant results for specifications for iBEATs and iBREATH.
SpecificationiBEATsiBREATH
Outlier rejection128, only async
330, both		111 only async
158 both
SD outlier rejection107, no rejection
107, 2SD
120, 2.5SD
124, 3SD		59 no rejection
70 2SD
104 2.5SD
36 3SD
Artifact trial rejection224, 85% criterion
162, small artifacts included
72, strict rejection		207, large artifacts included
62, small artifacts included
0, strict rejection
Data transformation240, log transformed
218, not transformed		162 log transformed
102 not transformed
Trial removal308, 0LTs included
150, 0LTs excluded		183, 0LTs included
86, 0LTs excluded
Min. number of trials per id to be included115, min. 2 trials
115, min. 4 trials
113, min. 8 trials
115, no criterion		86, min. 2 trials
74, min. 4 trials
24, min. 8 trials
85, no criterion
Statistical analysis220, linear mixed model
238, paired t-test		89, linear mixed model
180, paired t-test

  1. Number of significant specifications for iBEATs and iBREATH separate for each category. Overall, there were 1024 specifications for iBEATs and 1536 for iBREATH. Detailed information on the specifications can be found in Appendix 4. Category 3 (artifact trial rejection) for the iBREATH was simplified to make comparison to the iBEATs more intuitive.

Table 7
Descriptive information for number of trials completed and included.
Paradigm, Age groupMcompletedSDcompletedMincludedSDincluded
iBEATs, 3 mo13.977.089.827.44
iBEATs, 9 mo18.166.359.526.63
iBEATs, 18 mo15.626.3110.907.54
iBREATH, 3 mo13.005.849.166.52
iBREATH, 9 mo13.254.8510.105.21
iBREATH, 18 mo12.527.636.885.56
Appendix 1—table 1
Main analysis for the 9-month-olds for iBEATs and iBREATH split by randomization.
ExperimentRandomizationNMean LT Synch (SD)Mean LT Async (SD)Comparison
iBEATs1466792 (3473)5727 (2508)t(45)=2.37, p=0.022
iBEATs268775 (5810)3734 (1175)V=21, p=0.031
iBREATH1536307 (3090)5563 (2282)t(52)=2.37, p=0.021
iBREATH236847 (1321)4085 (331)V=6, p=0.25

  1. LT = looking time, Synch = synchronous trials, Async = asynchronous trials. For the larger samples a paired t-test was used, and a Wilcoxon rank sign test for the smaller samples.

Appendix 1—table 2
Mean looking time for first and second half of trials in the iBEATs.
AgeHalfMean looking time (SD) synchronousMean looking time (SD) asynchronous
3 mo17620 (4786)6686 (6193)
3 mo26010 (5899)5306 (5443)
9 mo18933 (4890)8197 (4954)
9 mo25724 (4475)5089 (3992)
18 mo18552 (5536)8575 (5772)
18 mo25607 (4527)5402 (4786)
Appendix 1—table 3
Mean looking time for first and second half of trials in the iBREATH.
AgeHalfMean looking time (SD) synchronousMean looking time (SD) asynchronous
3 mos19188 (9176)8356 (8654)
3 mo25903 (7876)6353 (7594)
9 mo18272 (5597)7110 (5486)
9 mo25774 (5425)5135 (4418)
18 mo16659 (5578)6976 (6629)
18 mo23550 (4124)4532 (3482)
Appendix 1—table 4
Mean looking times (SD) in ms and number of trials for side of the screen (left/right), condition, and age groups.
AgeSideConditionNumber of trials iBEATsMean looking time (SD) in ms iBEATsNumber of trials iBREATHMean looking time (SD) in ms iBREATH
3LSynchronous1165807 (5854)957140 (8689)
LAsynchronous1855708 (5835)1357138 (8230)
RSynchronous1817367 (6130)1477841 (8673)
RAsynchronous1175978 (5775)817461 (7806)
9LSynchronous1587137 (5066)876948 (6137)
LAsynchronous2386962 (4963)1976185 (5418)
RSynchronous2427451 (4889)1896567 (5235)
RAsynchronous1735619 (4269)1075642 (4416)
18LSynchronous697206 (5323)235099 (5157)
LAsynchronous1086861 (5395)785433 (5567)
RSynchronous1096867 (5286)825104 (5076)
RAsynchronous676904 (5753)415694 (4890)

  1. Values displayed here are based on the pre-processing pipeline used for the MEGA-analysis and are computed with unaggregated dataset. L=left, R=right.

Appendix 3—table 1
Full-null model comparison for the iBEATs model.
ModelDfAICBIClogLikDevianceChisqChi Dfp-value
Null5–1222.9–1195.5616.46–1232.9   
Full8–1227.8–1184.0621.92–1243.810.9130.012
Appendix 3—table 2
Results for the MEGA analysis of the iBEATs data with 3 mo as reference group.
TermEstimateSEz-valuep-value
Intercept–1.080.12–8.69<0.001
Condition asynchronous–0.170.08–2.150.031
9 mo0.570.163.48<0.001
18 mo0.440.182.490.013
Condition * 9 mo0.000.110.020.982
Condition * 18 mo0.140.131.080.283

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 3 mo age group.

Appendix 3—table 3
Results for the MEGA analysis of the iBEATs data with 9 mo as reference group.
TermEstimateSEz-valuep-value
Intercept–0.500.11–4.66<0.001
Condition asynchronous–0.170.07–2.490.013
3 mo–0.570.16–3.48<0.001
18 mo–0.130.10–1.220.221
Condition * 3 mo–0.000.11–0.020.982
Condition * 18 mo0.140.121.120.264

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 9 mo age group.

Appendix 3—table 4
Results for the MEGA analysis of the iBEATs data with 18 mo as reference group.
TermEstimateSEz-valuep-value
Intercept–0.630.13–4.91<0.001
Condition asynchronous–0.030.10–0.310.756
3 mo0.440.182.490.013
9 mo–0.130.10–1.220.221
Condition * 3 mo–0.140.13–1.080.283
Condition * 9 mo–0.140.12–1.120.264

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 18 mo age group.

Appendix 3—table 5
Full-null model comparison for the iBREATH model.
ModelDfAICBIClogLikDevianceChisqChi Dfp-value
Null5–1600.3–1574.5805.14–1610.3
Full8–1600.7–1559.4808.37–1616.76.4530.091
Appendix 3—table 6
Results for the MEGA analysis of the iBREATH data.
TermEstimateSEz-valuep-value
Intercept–1.350.13–10.35<0.001
Condition asynchronous–0.150.09–1.740.082
9 mo0.250.171.470.141
18 mo–0.150.19–0.770.440
Condition * 9 mo0.020.120.170.864
Condition * 18 mo0.230.161.500.134

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 3 mo age group.

Appendix 3—table 7
Results for the MEGA analysis of the iBREATH data.
TermEstimateSEz-valuep-value
Intercept–1.100.11–9.76<0.001
Condition asynchronous–0.130.08–1.750.080
3 mo–0.250.17–1.470.141
18 mo–0.400.13–3.120.001
Condition * 3 mo–0.020.12–0.170.864
Condition * 18 mo0.210.151.430.154

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 9 mo age group.

Appendix 3—table 8
Results for the MEGA analysis of the iBREATH data.
TermEstimateSEz-valuep-value
Intercept–1.350.13–10.35<0.001
Condition asynchronous–0.150.09–1.740.082
3 mo0.250.171.470.141
9 mos–0.150.19–0.770.440
Condition * 3 mo–0.230.16–1.500.134
Condition * 9 mo–0.210.15–1.430.154

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the synchronous condition, and 18 mo age group.

Appendix 3—table 9
MEGA analysis for the relationship between iBEATs and iBREATH with 3 mo as reference group.
TermEstimateSEz-valuep-value
Intercept–1.170.22–5.25<0.001
iBEATs score–1.830.97–1.890.059
9 mo–0.150.34–0.420.674
18 mo–0.050.35–0.150.880
iBEATs * 9 mo0.671.310.510.610
iBEATs * 18 mo3.131.412.220.027

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the 3 mo age group.

Appendix 3—table 10
MEGA analysis for the relationship between iBEATs and iBREATH with 9 mo as reference group.
TermEstimateSEz-valuep-value
Intercept–1.130.27–4.89<0.001
iBEATs score–1.160.90–1.300.192
3 mo0.150.340.420.674
18 mo0.090.380.240.810
iBEATs * 3 mo–0.671.31–0.510.610
iBEATs * 18 mo2.451.361.810.070

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the 9 mo age group.

Appendix 3—table 11
MEGA analysis for the relationship between iBEATs and iBREATH with 18 mo as reference group.
TermEstimateSEz-valuep-value
Intercept–1.220.28–4.39<.001
iBEATs score1.301.021.27.204
3 mo0.050.350.15.880
9 mos–0.090.38–0.24.810
iBEATs * 3 mo–3.131.41–2.22.027
iBEATs * 9 mo–2.451.36–1.81.070

  1. Results for the mixed model using a beta error distribution. Results are in comparison to the 18 mo age group.

Appendix 4—table 1
Analytical decisions for the iBEATs.
CategoryImplementation specification curve analysisNumber of analytical choices
Outlier rejectionOnly for sync trials, apply to all trials2
SD outlier rejection2SD, 2.5SD, 3SD, no criterion4
ECG artifact trial rejectionAll R-peaks identified, missed single R-peaks, missed R-peaks in last two seconds, identified at least 85% of R-peaks4
Data transformationLog-transformation, non-transformed data2
Trial removalRemove trials with 0 looking times, keep trials with 0 looking times2
Min. number of trials per id to be included8, 4, 2, 04
Statistical analysisPaired t-test, linear mixed model (trials clustered in id)2
Appendix 4—table 2
Analytical decisions for the iBREATH.
CategoryImplementation specification curve analysisNumber of analytical choices
Outlier rejectionOnly for sync trials, apply to all trials2
SD outlier rejection2SD, 2.5SD, 3SD, no criterion4
Respiration artifact trial rejectionOnly good signals, include deep breaths, artifacts in ECG but not respiration, deep breaths & artifacts in ECG, short flat lines in respiratory signal, flat lines & deep breaths and/or ECG artifacts6
Data transformationLog-transformation, non-transformed data2
Trial removalRemove trials with 0 looking times, keep trials with 0 looking times2
Min. number of trials per id to be included8, 4, 2, 04
Statistical analysisPaired t-test, linear mixed model (trials clustered in id)2
Appendix 4—table 3
Inference of the specification curve analysis.
Test statistic usedObserved resultp-value (% of shuffled results as or more extreme than observed results)
iBEATs
1. Median effect size0.1110.008
2. Share of significant results4700.042
3. Aggregate all p-valuesStouffer Z=–39.910.068
iBREATH
1. Median effect size0.1090.032
2. Share of significant results2690.088
3. Aggregate all p-valuesStouffer Z=–24.320.120
Appendix 5—table 1
Looking times for slow and fast asynchronous trials.
Mean Async (SD) in msMean Slow (SD) in msMean Fast (SD) in msComparison
iBEATs, 3 mo5812 (5804)5652
(5731)		5961
(5885)		t(299.15)=0.46,
p=0.64
iBEATs, 9 mo6539 (4924)6517
(5125)		6655
(4755)		t(536.77)=0.33,
p=0.74
iBEATs, 18 mo6877 (5518)6608
(5473)		7120
(5578)		t(171.77)=0.61,
p=0.54
iBREATH, 3 mo7259 (8057)7924
(8601)		6630
(7491)		t(206.34)=–1.18,
p=0.24
iBREATH, 9 mo6123 (5282)6657
(5779)		5659
(4939)		t(366.26)=–1.85,
p=0.066
iBREATH, 18 mo5477 (5325)6078
(6053)		4969
(4610)		t(99.74)=–1.11,
p=0.27

  1. For this analysis we used the pre-processing of the MEGA-analysis. Thus, looking times might differ from the confirmatory analysis.

Additional files

MDAR checklist
https://cdn.elifesciences.org/articles/91579/elife-91579-mdarchecklist1-v1.docx

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Markus R Tünte
  2. Stefanie Hoehl
  3. Moritz Wunderwald
  4. Johannes Bullinger
  5. Asena Boyadziheva
  6. Lara Maister
  7. Birgit Elsner
  8. Manos Tsakiris
  9. Ezgi Kayhan
(2025)
Respiratory and cardiac interoceptive sensitivity in the first two years of life
eLife 12:RP91579.
https://doi.org/10.7554/eLife.91579.4