Reconstruction algorithm for frequency-differential EIT using absolute values

Abstract

The human body consists of various biological tissues with frequency-dependent electrical conductivities. These properties change over time, for example due to cardio-respiratory activity. Electrical impedance tomography (EIT) injects a harmless current into the body and measures the resulting surface potential to recover regional conductivity distributions inside the human body. Time-differential EIT (tdEIT) provides information about conductivity changes over time. tdEIT assumes that the change in conductivity is small compared to the reference measurement and is only local, so that the background conductivity remains constant. This assumption allows a linear reconstruction, e.g. using the GREIT algorithm (Adler et al 2009). This technique can be used to monitor ventilation of critically ill patients, so that most EIT-applications address breath-by-breath lung monitoring using tdEIT approaches. However, many lung pathologies develop over a long period of time and thus can not provide a reference image for tdEIT. An edema, for example, impairs the gas exchange due to fluid accumulation in the airspace of the lung. Similar symptoms arise due to a pneumonia, which is an inflammatory infection of the alveoli. However, multi-frequency EIT (mfEIT) might have the potential to monitor lung pathologies without a baseline (Packham et al 2012). The main fundamental characteristic of multi-frequency EIT (mfEIT) is the combination of tissue properties and regional information to reconstruct the location of various tissues inside the body. The reconstruction of mfEIT is much more complex than tdEIT, as mfEIT performs simultaneous measurements at multiple frequencies, which contain the regional frequency-dependent conductivity. This was utilized in Mayer et al (2006), who reported a method for direct reconstruction of tissue-parameter. A similar principle was later specified by Malone et al (2013). This spectral constraints algorithm (SCA) uses multiple frequencies to reconstruct a tissue fraction inside the image, which can be described as a tissue-probability map. The number of constraints in the reconstruction is increased due to prior knowledge about the underlying

Efficient Multitask Structure-Aware Sparse Bayesian Learning for Frequency-Difference Electrical Impedance Tomography

“…Additionally, salt solution provides a homogeneous background so that the performance of different algorithms can be directly compared and the modelling errors are negligible. It is noteworthy that complex background substances such as cucumber, potato, and banana mashes have also been utilized in some studies [28], [29]. Those backgrounds exhibit reasonable permittivities and better simulate certain realistic biomedical scenarios.…”

Time Sequence Learning for Electrical Impedance Tomography Using Bayesian Spatiotemporal Priors

“…Additionally, salt solution provides a homogeneous background so that the performance of different algorithms can be directly compared and the modelling errors are negligible. It is noteworthy that complex background substances such as cucumber, potato, and banana mashes have also been utilized in some studies [28], [29]. Those backgrounds exhibit reasonable permittivities…”

Multi-channel bioimpedance spectroscopy based on orthogonal baseband shifting

“…The OBS method overcomes the drawback of channel-wise ADC and computational intensive demodulation by measuring all of the resulting voltages at once. Besides, multi-frequency EIT measures simultaneously at different frequencies and has the potential for tissue identification based on regional spectral impedance distributions (Menden et al 2019). OBS is an ideal principle for multi-spectral EIT acquisition at very high frame rates.…”