Detection and evaluation of bursts in terms of novelty and surprise
The detection of bursts and also of response onsets is often of relevance in understanding neurophysiological data, but the detection of these events is not a trivial task. We build on a method that was originally designed for burst detection using the so-called burst surprise as a measure. We exten...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIMS Press
2019-07-01
|
| Series: | Mathematical Biosciences and Engineering |
| Subjects: | |
| Online Access: | https://www.aimspress.com/article/10.3934/mbe.2019351?viewType=HTML |
| _version_ | 1828398823340244992 |
|---|---|
| author | Junji Ito Junji Ito Emanuele Lucrezia Günther Palm Sonja Grün |
| author_facet | Junji Ito Junji Ito Emanuele Lucrezia Günther Palm Sonja Grün |
| author_sort | Junji Ito |
| collection | DOAJ |
| description | The detection of bursts and also of response onsets is often of relevance in understanding neurophysiological data, but the detection of these events is not a trivial task. We build on a method that was originally designed for burst detection using the so-called burst surprise as a measure. We extend this method and provide a proper significance measure. Our method consists of two stages. In the first stage we model the neuron's interspike interval (ISI) distribution and make an i.i.d. assumption to formulate our null hypothesis. In addition we define a set of 'surprising' events that signify deviations from the null hypothesis in the direction of 'burstiness'. Here the so-called (strict) burst novelty is used to measure the size of this deviation. In the second stage we determine the significance of this deviation. The (strict) burst surprise is used to measure the significance, since it is the negative logarithm of the significance probability. After showing the consequences of a non-proper null hypothesis on burst detection performance, we apply the method to experimental data. For this application the data are divided into a period for parameter estimation to express a proper null hypothesis (model of the ISI distribution), and the rest of the data is analyzed by using that null hypothesis. We find that assuming a Poisson process for experimental spike data from motor cortex is rarely a proper null hypothesis, because these data tend to fire more regularly and thus a gamma process is more appropriate. We show that our burst detection method can be used for rate change onset detection, because a deviation from the null hypothesis detected by (strict) burst novelty also covers an increase of firing rate. |
| first_indexed | 2024-12-10T09:07:44Z |
| format | Article |
| id | doaj.art-81fde183e4fe4443acf01d0ad55cc83f |
| institution | Directory Open Access Journal |
| issn | 1551-0018 |
| language | English |
| last_indexed | 2024-12-10T09:07:44Z |
| publishDate | 2019-07-01 |
| publisher | AIMS Press |
| record_format | Article |
| series | Mathematical Biosciences and Engineering |
| spelling | doaj.art-81fde183e4fe4443acf01d0ad55cc83f2022-12-22T01:55:06ZengAIMS PressMathematical Biosciences and Engineering1551-00182019-07-011666990700810.3934/mbe.2019351Detection and evaluation of bursts in terms of novelty and surpriseJunji Ito0Junji Ito1Emanuele Lucrezia2Günther Palm3Sonja Grün41. Institute for Neuroscience and Medicine (INM-6), Jülich Research Centre, 52425 Jülich, Germany 2. Institute for Advanced Simulation (IAS-6) and Jara Brain Institute for Brain Structure and Function (INM-10), Jülich Research Centre, 52425 Jülich, Germany1. Institute for Neuroscience and Medicine (INM-6), Jülich Research Centre, 52425 Jülich, Germany 2. Institute for Advanced Simulation (IAS-6) and Jara Brain Institute for Brain Structure and Function (INM-10), Jülich Research Centre, 52425 Jülich, Germany1. Institute for Neuroscience and Medicine (INM-6), Jülich Research Centre, 52425 Jülich, Germany 2. Institute for Advanced Simulation (IAS-6) and Jara Brain Institute for Brain Structure and Function (INM-10), Jülich Research Centre, 52425 Jülich, Germany1. Institute for Neuroscience and Medicine (INM-6), Jülich Research Centre, 52425 Jülich, Germany3. Institute of Neural Information Processing, University of Ulm, James-Franck-Ring, 89081 Ulm, Germany1. Institute for Neuroscience and Medicine (INM-6), Jülich Research Centre, 52425 Jülich, Germany 2. Institute for Advanced Simulation (IAS-6) and Jara Brain Institute for Brain Structure and Function (INM-10), Jülich Research Centre, 52425 Jülich, Germany4. Theoretical Systems Neurobiology, RWTH Aachen University, Worringerweg 3, 52056 Aachen, GermanyThe detection of bursts and also of response onsets is often of relevance in understanding neurophysiological data, but the detection of these events is not a trivial task. We build on a method that was originally designed for burst detection using the so-called burst surprise as a measure. We extend this method and provide a proper significance measure. Our method consists of two stages. In the first stage we model the neuron's interspike interval (ISI) distribution and make an i.i.d. assumption to formulate our null hypothesis. In addition we define a set of 'surprising' events that signify deviations from the null hypothesis in the direction of 'burstiness'. Here the so-called (strict) burst novelty is used to measure the size of this deviation. In the second stage we determine the significance of this deviation. The (strict) burst surprise is used to measure the significance, since it is the negative logarithm of the significance probability. After showing the consequences of a non-proper null hypothesis on burst detection performance, we apply the method to experimental data. For this application the data are divided into a period for parameter estimation to express a proper null hypothesis (model of the ISI distribution), and the rest of the data is analyzed by using that null hypothesis. We find that assuming a Poisson process for experimental spike data from motor cortex is rarely a proper null hypothesis, because these data tend to fire more regularly and thus a gamma process is more appropriate. We show that our burst detection method can be used for rate change onset detection, because a deviation from the null hypothesis detected by (strict) burst novelty also covers an increase of firing rate.https://www.aimspress.com/article/10.3934/mbe.2019351?viewType=HTMLburst detectioninterspike intervalssignificancegamma processresponse onset detection |
| spellingShingle | Junji Ito Junji Ito Emanuele Lucrezia Günther Palm Sonja Grün Detection and evaluation of bursts in terms of novelty and surprise Mathematical Biosciences and Engineering burst detection interspike intervals significance gamma process response onset detection |
| title | Detection and evaluation of bursts in terms of novelty and surprise |
| title_full | Detection and evaluation of bursts in terms of novelty and surprise |
| title_fullStr | Detection and evaluation of bursts in terms of novelty and surprise |
| title_full_unstemmed | Detection and evaluation of bursts in terms of novelty and surprise |
| title_short | Detection and evaluation of bursts in terms of novelty and surprise |
| title_sort | detection and evaluation of bursts in terms of novelty and surprise |
| topic | burst detection interspike intervals significance gamma process response onset detection |
| url | https://www.aimspress.com/article/10.3934/mbe.2019351?viewType=HTML |
| work_keys_str_mv | AT junjiito detectionandevaluationofburstsintermsofnoveltyandsurprise AT junjiito detectionandevaluationofburstsintermsofnoveltyandsurprise AT emanuelelucrezia detectionandevaluationofburstsintermsofnoveltyandsurprise AT guntherpalm detectionandevaluationofburstsintermsofnoveltyandsurprise AT sonjagrun detectionandevaluationofburstsintermsofnoveltyandsurprise |