<i>catch22</i>: CAnonical Time-series CHaracteristics <i>selected through</i> highly comparative time-series analysis

Lubba, Carl Henning; Sethi, Sarab S.; Knaute, Philip; Schultz, Simon R.; Fulcher, Ben D.; Jones, Nick

doi:10.1101/532259

Cited by 10 publications

(9 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Longest period below mean (LPBM) LPBM measures the time length of the longest inactive periods (or active periods depending on the time series characteristics) in a given time series (Lubba et al, 2019); first, the time lengths of all consecutive values below the mean of time series are calculated and then the maximum of the time lengths are obtained as a LPBM value. LPBM is known to be one of the important temporal statistics in time series analysis (Lubba et al, 2019), and was used in this study to be a measure of persistent inactiveness of spike activity. A high LPBM value in MUA signal implies an existence of long inactive period suggesting synchronous fragmentation of spike activities, whereas a low LPBM indicates more continuous activity suggesting an irregular and asynchronous spiking pattern.…”

Section: Multi-unit Spike Propertiesmentioning

confidence: 99%

State-Dependent Cortical Unit Activity Reflects Dynamic Brain State Transitions in Anesthesia

2020

View full text Add to dashboard Cite

How anesthesia affects cortical neuronal spiking and information transfer could help understand the neuronal basis of conscious state. Recent investigations suggest that global state of the anesthetized brain is not stationary but changes spontaneously at a fixed level of anesthetic concentration. How cortical unit activity changes with dynamically transitioning brain states under anesthesia is unclear. We hypothesized that distinct cortical states are characterized by distinct neuronal spike patterns.Extracellular unit activity was measured with sixty-four-channel silicon microelectrode arrays in cortical layers 5/6 of primary visual cortex of chronically instrumented, freely moving male rats (N = 7) during stepwise reduction of the anesthetic desflurane (6, 4, 2, and 0%). Unsupervised machine learning applied to multi-unit spike patterns revealed five distinct brain states of which four occurred at various anesthetic concentrations and shifted spontaneously. In deeper anesthesia states, the number of active units and overall spike rate decreased while the remaining active units showed increased bursting (excitatory neurons), spike timing variability, unit-to-population correlation and unit-to-unit transfer entropy, especially among putative excitatory units, despite the overall decrease in transfer entropy. A novel desynchronized brain state with increased spike timing variability, entropy and electromyographic activity that occurred mostly in deep anesthesia was discovered. These results provide evidence for distinct unit activity patterns associated with spontaneous changes in local cortical brain states at stationary anesthetic conditions. The appearance of a paradoxical, desynchronized brain state in deep anesthesia contends the prevailing view of monotonic dose-dependent anesthetic effects on the brain.. CC-BY-NC-ND 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made SIGNIFICANCE STATEMENTRecent studies suggest that spontaneous changes in brain state occur under anesthesia. However, the spiking behavior of cortical neurons associated with such state changes has not been investigated. We found that local brain states defined by multi-unit activity had non-unitary relationship with the current anesthetic level. A paradoxical brain state displaying asynchronous firing pattern and high electromyographic activity was found unexpectedly at high-dose anesthesia. In contrast, the synchronous fragmentation of neuronal spiking appeared to be a robust signature of the state of anesthesia. The findings challenge the assumption of monotonic, anesthetic dose-dependent behavior of cortical neuron populations. They enhance the interpretation of neuroscientific data obtained under anesthesia and understanding of the neuronal basis of anesthetic-induced state of unconsciousness..

show abstract

Section: Multi-unit Spike Propertiesmentioning

confidence: 99%

State-Dependent Cortical Unit Activity Reflects Dynamic Brain State Transitions in Anesthesia

2020

View full text Add to dashboard Cite

show abstract

“…Our modeling approach is based on a set of wellestablished methods to reduce the dimensionality of PROST. While our results are in favor of a data-driven variable selection, there is room to experiment further with different time-series characterizations (e.g., Lubba et al, 2019). It is also possible to expand our framework to a more granular level, for example, using regional PROST for better planning of the distribution.…”

Section: Model Extensionsmentioning

confidence: 89%

Predictive competitive intelligence with prerelease online search traffic

Schaer

Kourentzes

Fildes

2022

Production and Operations Management

View full text Add to dashboard Cite

In today's competitive market environment, it is vital for companies to gain insight about competitors' new product launches. Past studies have demonstrated the predictive value of prerelease online search traffic (PROST) for new product forecasting. Relying on these findings and the public availability of PROST, we investigate its usefulness for estimating sales of competing products. We propose a model for predicting the success of competitors' product launches, based on own past product sales data and competitor's prerelease Google Trends. We find that PROST increases predictive accuracy by more than 18% compared to models that only use internally available sales data and product characteristics of video game sales. We conclude that this inexpensive source of competitive intelligence can be helpful when managing the marketing mix and planning new product releases.

show abstract

“…Although the relevant features from physiological signals can be extracted manually using pre-defined patterns described in medical literature, we tried to automate the feature extraction process by simply considering the physiological signals as time-series data. We extracted 22 time-series features reported to be most effective by Lubba et al [ 51 ]. They selected these 22 features by exploring about 5000 candidate features for their classification performance on 92 different time-series datasets.…”

Section: Methodsmentioning

confidence: 99%

Exploration of physiological sensors, features, and machine learning models for pain intensity estimation

2021

View full text Add to dashboard Cite

In current clinical settings, typically pain is measured by a patient’s self-reported information. This subjective pain assessment results in suboptimal treatment plans, over-prescription of opioids, and drug-seeking behavior among patients. In the present study, we explored automatic objective pain intensity estimation machine learning models using inputs from physiological sensors. This study uses BioVid Heat Pain Dataset. We extracted features from Electrodermal Activity (EDA), Electrocardiogram (ECG), Electromyogram (EMG) signals collected from study participants subjected to heat pain. We built different machine learning models, including Linear Regression, Support Vector Regression (SVR), Neural Networks and Extreme Gradient Boosting for continuous value pain intensity estimation. Then we identified the physiological sensor, feature set and machine learning model that give the best predictive performance. We found that EDA is the most information-rich sensor for continuous pain intensity prediction. A set of only 3 features from EDA signals using SVR model gave an average performance of 0.93 mean absolute error (MAE) and 1.16 root means square error (RMSE) for the subject-independent model and of 0.92 MAE and 1.13 RMSE for subject-dependent. The MAE achieved with signal-feature-model combination is less than 1 unit on 0 to 4 continues pain scale, which is smaller than the MAE achieved by the methods reported in the literature. These results demonstrate that it is possible to estimate pain intensity of a patient using a computationally inexpensive machine learning model with 3 statistical features from EDA signal which can be collected from a wrist biosensor. This method paves a way to developing a wearable pain measurement device.

show abstract

catch22: CAnonical Time-series CHaracteristics selected through highly comparative time-series analysis

Cited by 10 publications

References 30 publications

State-Dependent Cortical Unit Activity Reflects Dynamic Brain State Transitions in Anesthesia

State-Dependent Cortical Unit Activity Reflects Dynamic Brain State Transitions in Anesthesia

Predictive competitive intelligence with prerelease online search traffic

Exploration of physiological sensors, features, and machine learning models for pain intensity estimation

Contact Info

Product

Resources

About