Process comments on v1 of JOSS paper

Author: psomhorst

paper/paper.md

@@ -1,5 +1,5 @@
 ---
-title: 'eitprocessing: a Python package for analysis of Electrial Impedance Tomography data'
+title: 'eitprocessing: a Python package for analysis of Electrical Impedance Tomography data'
 tags:
   - Python
   - Electrical Impedance Tomography
@@ -8,25 +8,25 @@ authors:
     orcid: 0000-0003-3490-2080
     equal-contrib: true
     affiliation: 1
-  - name: Dani Bodor
+  - name: Dani L. Bodor
     equal-contrib: true
     orcid: 0000-0003-2109-2349
     affiliation: 2
   - name: Juliette Francovich
     affiliation: 1
-    orcid: 0000-0000-0000-0000
-  - name: Jantine Wisse-Smit
+    orcid: 0009-0004-0976-5082
+  - name: Jantine J. Wisse
     affiliation: 1
-    orcid: 0000-0000-0000-0000
+    orcid: 0009-0006-6552-5459
   - name: Walter Baccinelli
     affiliation: 2
     orcid: 0000-0001-8888-4792
-  - name: Annemijn Jonkman
+  - name: Annemijn H. Jonkman
     orcid: 0000-0001-8778-5135
     affiliation: 1
 
 affiliations:
- - name: Department of Intensive Care adults, Erasmus MC, Rotterdam, The Netherlands
+ - name: Department of Adult Intensive Care, Erasmus MC, Rotterdam, The Netherlands
    index: 1
  - name: Netherlands eScience Center, Amsterdam, The Netherlands 
    index: 2
@@ -37,96 +37,101 @@ bibliography: paper.bib
 ## Summary
 
 Electrical Impedance Tomography (EIT) is a promising non-invasive, radiation-free technology for
-monitoring the respiratory system of patients who undergo mechanical ventilation in the Intensive
+monitoring the respiratory system. EIT is mostly used to optimize ventilator settings to the respiratory mechanics of
+mechanically ventilated patients in the Intensive
 Care Unit. While EIT is gaining popularity, the complexity of data processing, analysis and
-interpretation hamper standardization, validation and widespread adoption. Commercial software is
-closed and opaque, while custom research software is often ad-hoc, single use and unverified.
+interpretation hampers standardization, validation and widespread adoption. Commercial software is
+closed and opaque, while custom research software is often ad-hoc, single use, and unverified.
 `eitprocessing` offers a standardized, open, and highly expandable pipeline for the processing and
-analysis of EIT and related data.
+analysis of EIT and respiration related data. 
 
-## Statement of need
+## State of the field
 
 Acute respiratory failure is the most common reason for admission to the intensive care unit (ICU),
 and can be caused by e.g., infection, trauma, heart failure, or complications during elective
 surgery. Patients with severely injured lungs and critically low levels of arterial oxygen require
 life-saving breathing support with mechanical ventilation [@tobin1998]. Although mechanical
-ventilation is the cornerstone of supportive therapy in the ICU, it is a double-edge sword:
+ventilation is the cornerstone of supportive therapy in the ICU, it is a double-edged sword:
 inadequate mechanical ventilator assist exacerbates lung injury and inflammation, and worsens
 outcomes [@Slutsky2013;@Amato2015]. ICU mortality for patients with acute respiratory failure
 remains high [~40%, @Bellani2016]; these numbers increased drastically during the COVID-19
 pandemic. To ameliorate the risk of death and long-term morbidity of the critically ill, we need
 mechanical ventilation strategies that are lung-protective and tailored to the individual patient’s
-respiratory physiology [@Goligher2020;@Goligher2020b]. However, there are yet no simple and
-reliable and readily accessible techniques available to clinicians at the bedside to identify the
+respiratory physiology [@Goligher2020;@Goligher2020b]. However, there are currently no simple,
+reliable, and readily accessible tools available to clinicians at the bedside to identify the
 beneficial and harmful effects of adaptations in mechanical ventilator support [@Jonkman2022].
 
-A very promising technology to change clinical practice in ICU patients is Electrical Impedance
-Tomography (EIT) [@Frerichs2016]. EIT is gaining popularity worldwide as a bedside non-invasive
-radiation-free lung imaging tool: using a belt mounted with electrodes placed around the chest, it
-continuously and real-time visualizes changes in lung volume owing to adaptations in ventilator
-pressures or because of changes in lung characteristics, resulting from worsening or improving lung
-function. In contrast to static anatomical imaging techniques such as computed tomography scan, EIT
+A very promising technology to change clinical practice in ICU patients is EIT [@Frerichs2016]. EIT is gaining worldwide popularity as a bedside non-invasive
+radiation-free tool for lung imaging. Using a belt fitted with electrodes placed around the chest, it
+continuously visualizes real-time changes in lung volume. These changes reflect tidal ventilation,
+changes in lung volume due to ventilator settings, and adaptations due variations in lung
+characteristics caused by improved or worsening lung mechanics. In contrast to static anatomical imaging techniques such as computed tomography
+scan, EIT
 provides dynamic information on lung ventilation. As such, EIT can monitor at the bedside the
-direct impact of mechanical ventilation on the lung and provides important information to assist in
-the clinical decision-making. Personalizing mechanical ventilation using EIT monitoring and
-diagnostics may ameliorate the risk of death and long-term morbidity, and may substantially reduce
-the burden on our healthcare system.
+direct impact of mechanical ventilation on the lung, help with personalizing mechanical
+ventilation, and assist in clinical decision-making. 
+Personalizing mechanical ventilation using EIT monitoring and diagnostics may ameliorate the risk
+of death and long-term morbidity, and may substantially reduce the burden on our healthcare system.
+
+## Statement of need
 
 The perspective that EIT will become an important standard monitoring technique is shared by
 international experts [@Frerichs2016;@Wisse2024-sl]. Both @Frerichs2016 and @Wisse2024-sl emphasize
-the importance of standardized techniques, terminology, and consensus regarding applications was
-extensively discussed. However, validated methods to implement EIT information in routine care are
-still lacking and synchronizing EIT data with other bedside respiratory monitoring modalities is
-often practically impossible. Standardized implementation of EIT information is further limited
-because the availability of bedside analysis tools depends on the type of EIT device used. Advanced
+the importance of standardized techniques, terminology, and consensus regarding the application of
+EIT. 
+Validated methods to implement EIT information in routine care are
+still lacking. Standardized implementation of EIT information is further limited
+as the availability of both bedside and offline analysis tools depends on the type of EIT device used. Advanced
 image and signal analysis could overcome certain challenges but also requires complex
-post-processing (including detection/removal of common artifacts) and specific technical expertise
+post-processing (including detection/removal of common artifacts) that is time-consuming and requires specific technical expertise
 that is often not present in clinical practice. This currently hampers reproducibility of research
-findings and clinical implementation. Moreover, it stresses the importance of close collaboration
+findings and clinical implementation. The current limitations of EIT analysis stresses the importance of close collaboration
 between physicians, clinical researchers and engineers in order to identify clinical needs, to
 develop and validate new algorithms, and to facilitate clinical implementation [@Scaramuzzo2024-ob].
 
 `eitprocessing` offers a standardized, open, and highly expandable library of tools for loading,
 filtering, segmentation and analysis of EIT data as well as related waveform or sparse data.
-`eitprocessing` can work with data from the three main clinically available EIT devices, as well as
-from related data sources. It includes commonly used methods for filtering and segmentation. The
-authors continuously develop further algorithms for analysis for current and future projects. The
+`eitprocessing` is compatible with data from the three most-used clinically available EIT devices, as well as
+from related data sources, such as mechanical ventilators and dedicated pressure devices. It includes commonly used methods for filtering and segmentation. The
+authors continuously develop and implement further algorithms for analysis. The
 international community has been invited to use and contribute to the software.
 
 ## Key features
 
 `eitprocessing` aims to simplify and standardize loading, pre-processing, analysis and reporting
-while working with different respiration-related datasets. 
-Notebooks showing these features are available in the repository.
+or respiration-related datasets. 
+Notebooks demonstrating these features are available in the repository.
 
 ### Loading
 
 `eitprocessing` supports the loading of EIT data exported from the Dräger Pulmovista (`.bin`
-files), Timpel Enlight (`.csv` files) and Sentec LuMon (`.zri`) devices. Non-EIT data saved in the
-data files are also loaded.
+files), Timpel Enlight (`.txt` files) and Sentec LuMon (`.zri` files) devices. Non-EIT data saved
+in the data files are also loaded.
 
 ### Data containers
 
-The main data container in `eitprocessing` is the sequence. A sequence represents a single
-continuous measurement of data in a single subject, and can contain multiple datasets. Sequences
-can be sliced --- by time or index --- and concatenated. Contained datasets are sliced and
-concatenated accordingly.
-
-`eitprocessing` currently supports four types of dataset. Continuous data has one-dimensional data
-points at predictable intervals with a fixed sample frequency. Examples are airway pressure
-measured by a mechanical ventilator or a global impedance signal. EIT data is also continuous, but
-contains three-dimensional data points, (generally) 32 rows by 32 columns by time. Sparse data has
-one-dimensional data points at unpredictable intervals and no set sample frequency. An example is
-the tidal volume measured by a mechanical ventilator, registered at the end of each breath.
-Interval data has one-dimensional data points that are valid for a time interval. An example is the
-position of a subject, e.g., supine for the first part of a measurement and prone for the second
-part.
+The main data container in `eitprocessing` is the `Sequence`. A sequence represents a single
+continuous measurement of data in a single subject, and can contain data from different sources. Sequences
+can be sliced --- by time or index --- and concatenated. All data contained in the sequence are
+sliced and concatenated accordingly.
+
+`eitprocessing` currently supports four types of dataset. The most important type is `EITData`,
+which contains the electrical impedance of individual pixels as three-dimensional data --- (generally)
+32 rows by 32 columns over time. Each frame of 32 by 32 pixels represents the impedance in a
+transverse plane through the thorax at the corresponding time. `ContinuousData` has one-dimensional data points at
+predictable intervals with a fixed sample frequency. Examples are airway pressure measured by a
+mechanical ventilator or a global impedance signal.  `SparseData` has one-dimensional data points at
+unpredictable intervals and no set sample frequency. An example is the tidal volume measured by a
+mechanical ventilator, registered at the end of each breath. `IntervalData` has one-dimensional data
+points that are valid for a time interval. An example is the position of a subject, e.g., supine
+for the first part of a measurement and prone for the second part.
 
 ### Pre-processing
 
 `eitprocessing` currently has implementations for the following pre-processing steps:
 
-- high-pass, low-pass, band-pass of band-stop Butterworth filters;
+- calculation of the global or regional impedance as the sum of the impedance of all or a subset of pixels;
+- high-pass, low-pass, band-pass or band-stop Butterworth filters;
 - a moving averager using convolution with a given window;
 - automatic detection of the start, middle (end-inspiration) and end of breaths on a
   global/regional and pixel level.
@@ -140,6 +145,9 @@ part.
 
 ## Future perspective
 
+`eitprocessing` is ready for use in offline analysis of EIT and respiratory related data. Our team
+is actively working on expanding the features of the software.
+
 Several features are in active development. Examples are:
 
 - more advanced filtering methods, using a combination of Butterworth filters, empirical mode