0
views
0
recommends
+1 Recommend
0 collections
    0
    shares
      • Record: found
      • Abstract: not found
      • Article: not found

      BI - Directional long short-term memory for automatic detection of sleep apnea events based on single channel EEG signal

      , , , , ,
      Computers in Biology and Medicine
      Elsevier BV

      Read this article at

      ScienceOpenPublisherPubMed
      Bookmark
          There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.

          Related collections

          Most cited references32

          • Record: found
          • Abstract: found
          • Article: not found

          Long Short-Term Memory

          Learning to store information over extended time intervals by recurrent backpropagation takes a very long time, mostly because of insufficient, decaying error backflow. We briefly review Hochreiter's (1991) analysis of this problem, then address it by introducing a novel, efficient, gradient-based method called long short-term memory (LSTM). Truncating the gradient where this does not do harm, LSTM can learn to bridge minimal time lags in excess of 1000 discrete-time steps by enforcing constant error flow through constant error carousels within special units. Multiplicative gate units learn to open and close access to the constant error flow. LSTM is local in space and time; its computational complexity per time step and weight is O(1). Our experiments with artificial data involve local, distributed, real-valued, and noisy pattern representations. In comparisons with real-time recurrent learning, back propagation through time, recurrent cascade correlation, Elman nets, and neural sequence chunking, LSTM leads to many more successful runs, and learns much faster. LSTM also solves complex, artificial long-time-lag tasks that have never been solved by previous recurrent network algorithms.
            Bookmark
            • Record: found
            • Abstract: found
            • Article: not found

            Rules for scoring respiratory events in sleep: update of the 2007 AASM Manual for the Scoring of Sleep and Associated Events. Deliberations of the Sleep Apnea Definitions Task Force of the American Academy of Sleep Medicine.

            The American Academy of Sleep Medicine (AASM) Sleep Apnea Definitions Task Force reviewed the current rules for scoring respiratory events in the 2007 AASM Manual for the Scoring and Sleep and Associated Events to determine if revision was indicated. The goals of the task force were (1) to clarify and simplify the current scoring rules, (2) to review evidence for new monitoring technologies relevant to the scoring rules, and (3) to strive for greater concordance between adult and pediatric rules. The task force reviewed the evidence cited by the AASM systematic review of the reliability and validity of scoring respiratory events published in 2007 and relevant studies that have appeared in the literature since that publication. Given the limitations of the published evidence, a consensus process was used to formulate the majority of the task force recommendations concerning revisions.The task force made recommendations concerning recommended and alternative sensors for the detection of apnea and hypopnea to be used during diagnostic and positive airway pressure (PAP) titration polysomnography. An alternative sensor is used if the recommended sensor fails or the signal is inaccurate. The PAP device flow signal is the recommended sensor for the detection of apnea, hypopnea, and respiratory effort related arousals (RERAs) during PAP titration studies. Appropriate filter settings for recording (display) of the nasal pressure signal to facilitate visualization of inspiratory flattening are also specified. The respiratory inductance plethysmography (RIP) signals to be used as alternative sensors for apnea and hypopnea detection are specified. The task force reached consensus on use of the same sensors for adult and pediatric patients except for the following: (1) the end-tidal PCO(2) signal can be used as an alternative sensor for apnea detection in children only, and (2) polyvinylidene fluoride (PVDF) belts can be used to monitor respiratory effort (thoracoabdominal belts) and as an alternative sensor for detection of apnea and hypopnea (PVDFsum) only in adults.The task force recommends the following changes to the 2007 respiratory scoring rules. Apnea in adults is scored when there is a drop in the peak signal excursion by ≥ 90% of pre-event baseline using an oronasal thermal sensor (diagnostic study), PAP device flow (titration study), or an alternative apnea sensor, for ≥ 10 seconds. Hypopnea in adults is scored when the peak signal excursions drop by ≥ 30% of pre-event baseline using nasal pressure (diagnostic study), PAP device flow (titration study), or an alternative sensor, for ≥ 10 seconds in association with either ≥ 3% arterial oxygen desaturation or an arousal. Scoring a hypopnea as either obstructive or central is now listed as optional, and the recommended scoring rules are presented. In children an apnea is scored when peak signal excursions drop by ≥ 90% of pre-event baseline using an oronasal thermal sensor (diagnostic study), PAP device flow (titration study), or an alternative sensor; and the event meets duration and respiratory effort criteria for an obstructive, mixed, or central apnea. A central apnea is scored in children when the event meets criteria for an apnea, there is an absence of inspiratory effort throughout the event, and at least one of the following is met: (1) the event is ≥ 20 seconds in duration, (2) the event is associated with an arousal or ≥ 3% oxygen desaturation, (3) (infants under 1 year of age only) the event is associated with a decrease in heart rate to less than 50 beats per minute for at least 5 seconds or less than 60 beats per minute for 15 seconds. A hypopnea is scored in children when the peak signal excursions drop is ≥ 30% of pre-event baseline using nasal pressure (diagnostic study), PAP device flow (titration study), or an alternative sensor, for ≥ the duration of 2 breaths in association with either ≥ 3% oxygen desaturation or an arousal. In children and adults, surrogates of the arterial PCO(2) are the end-tidal PCO(2) or transcutaneous PCO(2) (diagnostic study) or transcutaneous PCO(2) (titration study). For adults, sleep hypoventilation is scored when the arterial PCO(2) (or surrogate) is > 55 mm Hg for ≥ 10 minutes or there is an increase in the arterial PCO(2) (or surrogate) ≥ 10 mm Hg (in comparison to an awake supine value) to a value exceeding 50 mm Hg for ≥ 10 minutes. For pediatric patients hypoventilation is scored when the arterial PCO(2) (or surrogate) is > 50 mm Hg for > 25% of total sleep time. In adults Cheyne-Stokes breathing is scored when both of the following are met: (1) there are episodes of ≥ 3 consecutive central apneas and/or central hypopneas separated by a crescendo and decrescendo change in breathing amplitude with a cycle length of at least 40 seconds (typically 45 to 90 seconds), and (2) there are five or more central apneas and/or central hypopneas per hour associated with the crescendo/decrescendo breathing pattern recorded over a minimum of 2 hours of monitoring.
              Bookmark
              • Record: found
              • Abstract: not found
              • Article: not found

              Bidirectional recurrent neural networks

                Bookmark

                Author and article information

                Contributors
                Journal
                Computers in Biology and Medicine
                Computers in Biology and Medicine
                Elsevier BV
                00104825
                March 2022
                March 2022
                : 142
                : 105211
                Article
                10.1016/j.compbiomed.2022.105211
                35007944
                6e5a82e0-6c67-4e07-950c-c317ec481ba1
                © 2022

                https://www.elsevier.com/tdm/userlicense/1.0/

                History

                Comments

                Comment on this article