Open Access
Add to BibSonomy
Abstract
In our original published article, we labeled the x-axis in Fig. 1(b) incorrectly [Biomed. Opt. Express 11, 1462 (2020) [CrossRef] ]. The sub-figure reports the importance of features extracted from the waveforms in training a machine learning algorithm to estimate intracranial pressure. This erratum corrects the labels in Fig. 1(b). The discussion and conclusions drawn from this article did not change.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
Results
Figure 1(b) [1] shows the importance of waveform features used in the machine learning algorithm. In the original article, the features in Fig. 1(b) were incorrectly labeled. This was due to an error in the label assignment inside the MATLAB script used to create this graph. After correcting this mistake, these features are now predominantly originating from the percussion wave (P1). The most important features, labeled MAP and AUC, for mean arterial pressure and area under the curve, are unaffected. The figure caption is also not affected.
This graphical mistake has no impact on the methods described in the original article nor does it impact the performance of the described algorithm in estimating intracranial pressure non-invasively from cardiac pulsations in blood flow measured from the brain with diffuse correlation spectroscopy.
Fig. 1. Results of the regression forest machine learning approach. (a) shows the distribution of the available data. The dashed line marks the maximum ICP level that was fitted for at 30 mmHg. (b) shows the distribution of features used in the regression forest as a percentage of all chosen features in all decision criteria generated. The standard deviation across individual trees is shown as error bars. Nomenclature is according to Table 1. (c) shows the performance of the regression forest by plotting estimated ICP (ICPest) over invasively measured ground truth (ICPinv). The solid line shows the ideal fit, while the dashed lines mark an area of 2 mmHg around the ideal fit. The shaded area shows the confidence interval. (d) graphs the difference between ICPest and ICPinv over ICPinv in a Bland-Altman plot. The dashed lines span a region of 95% of the distribution, corresponding to a standard deviation of 1.96. The histogram on the right of this graph shows the distribution of data points in number of samples. (e) shows a continuous estimate of ICP for NHP 3. The gray line shows the estimated ICP, and the black line the invasively measured ICP. An r2 = 0.92 and a mean squared error MSE = 3.2 mmHg2 were achieved.
Funding
Center for Machine Learning and Health at Carnegie Mellon University; American Heart Association (17SDG33700047); National Institutes of Health (R21-EB024675).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
References
1. A. Ruesch, J. Yang, S. Schmitt, D. Acharya, M. A. Smith, and J. M. Kainerstorfer, “Estimating intracranial pressure using pulsatile cerebral blood flow measured with diffuse correlation spectroscopy,” Biomed. Opt. Express 11(3), 1462 (2020). [CrossRef]
