Harriet Akre
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Harriet Akre has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Sigurd Steinshamn
Department of Circulation and Medical Imaging, Norwegian University of Science and Technology, Trondheim, Norway
4 shared publications - Britt Øverland
Pediatric and Adult Sleep Disorder Clinic, Lovisenberg Diaconal Hospital, Oslo, Norway
3 shared publications - Knut Liestøl
University of Oslo
3 shared publications - Konstantinos Nikolaidis
Department of Informatics, University of Oslo, Oslo, Norway
3 shared publications - Lars Aakerøy
Department of Thoracic Medicine St. Olavs Hospital, Trondheim University Hospital Trondheim Norway
3 shared publications - Mohan Kankanhalli
National University of Singapore, Singapore, Singapore
3 shared publications
Latest Publications
- Open Access
- 30 October 2024
- Published by BMJ in Open Heart
- Vol. 11 (2) , e002718
Cardiac remodelling in patients with atrial fibrillation and obstructive sleep apnoea
Background: Obstructive sleep apnoea (OSA) can cause left atrial (LA) and left ventricular (LV) remodelling, which is linked to atrial fibrillation (AF). Whether continuous positive airway pressure (CPAP) can reverse LA and LV remodelling in patients with OSA and paroxysmal AF (PAF) has yet to be studied. We assessed the impact of CPAP treatment on LA and LV size and function in patients with OSA and PAF before and after catheter ablation. Methods: In a randomised controlled trial, we screened patients with PAF for OSA. We enrolled patients with an Apnoea–Hypopnoea Index ≥15/hour. The burden of AF was monitored by an implantable loop recorder in all patients. Patients were then randomised to CPAP treatment or standard care. Transthoracic echocardiography was performed at baseline and after 6 and 12 months to assess LV and LA function and remodelling with advanced echocardiographic imaging techniques. Results: We enrolled 109 patients (63±7 years, body mass index 29.6±4.3, 76% men). 83 patients were scheduled for pulmonary vein isolation (PVI) and 26 for clinical follow-up only. 55 patients were randomised to CPAP and 54 to standard care. The burden of AF decreased significantly in patients who underwent PVI irrespective of treatment with CPAP (p for difference ≤0.001). Patients in the study group had LV ejection fraction (LVEF) and LV global longitudinal strain (GLS) within the normal range, increased LA Volume Index (LAVI), LA volume (by speckle tracking) and decreased LA reservoir strain at baseline. We did not observe any improvement in LVEF, GLS, LAVI, LA volume or LA reservoir strain in either group during the 12 months of follow-up. Conclusions: In patients with PAF and OSA, treatment with CPAP was not associated with reverse LA remodelling within 12 months of follow-up.
- 30 October 2024
- Published by BMJ in Open Heart
- Vol. 11 (2) , e002718
Open AccessCardiac remodelling in patients with atrial fibrillation and obstructive sleep apnoea
Background: Obstructive sleep apnoea (OSA) can cause left atrial (LA) and left ventricular (LV) remodelling, which is linked to atrial fibrillation (AF). Whether continuous positive airway pressure (CPAP) can reverse LA and LV remodelling in patients with OSA and paroxysmal AF (PAF) has yet to be studied. We assessed the impact of CPAP treatment on LA and LV size and function in patients with OSA and PAF before and after catheter ablation. Methods: In a randomised controlled trial, we screened patients with PAF for OSA. We enrolled patients with an Apnoea–Hypopnoea Index ≥15/hour. The burden of AF was monitored by an implantable loop recorder in all patients. Patients were then randomised to CPAP treatment or standard care. Transthoracic echocardiography was performed at baseline and after 6 and 12 months to assess LV and LA function and remodelling with advanced echocardiographic imaging techniques. Results: We enrolled 109 patients (63±7 years, body mass index 29.6±4.3, 76% men). 83 patients were scheduled for pulmonary vein isolation (PVI) and 26 for clinical follow-up only. 55 patients were randomised to CPAP and 54 to standard care. The burden of AF decreased significantly in patients who underwent PVI irrespective of treatment with CPAP (p for difference ≤0.001). Patients in the study group had LV ejection fraction (LVEF) and LV global longitudinal strain (GLS) within the normal range, increased LA Volume Index (LAVI), LA volume (by speckle tracking) and decreased LA reservoir strain at baseline. We did not observe any improvement in LVEF, GLS, LAVI, LA volume or LA reservoir strain in either group during the 12 months of follow-up. Conclusions: In patients with PAF and OSA, treatment with CPAP was not associated with reverse LA remodelling within 12 months of follow-up.
- Open Access
- 31 October 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
- Vol. 3 (4) , 1-24
My Health Sensor, My Classifier – Adapting a Trained Classifier to Unlabeled End-User Data
Sleep apnea is a common yet severely under-diagnosed sleep related disorder. Unattended sleep monitoring at home with low-cost sensors can be leveraged for condition detection, and Machine Learning offers a generalized solution for this task. However, patient characteristics, lack of sufficient training data, and other factors can imply a domain shift between training and end-user data and reduced task performance. In this work, we address this issue with the aim to achieve personalization based on the patient’s needs. We present an unsupervised domain adaptation (UDA) solution with the constraint that labeled source data are not directly available. Instead, a classifier trained on the source data is provided. Our solution iteratively labels target data sub-regions based on classifier beliefs, and trains new classifiers from the expanding dataset. Experiments with sleep monitoring datasets and various sensors show that our solution outperforms the classifier trained on the source domain, with a kappa coefficient improvement from 0.012 to 0.242. Additionally, we apply our solution to digit classification DA between three well-established datasets, to investigate its generalizability, and allow for related work comparisons. Even without direct access to the source data, it outperforms several well-established UDA methods in these datasets.
- 31 October 2022
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
- Vol. 3 (4) , 1-24
Open AccessMy Health Sensor, My Classifier – Adapting a Trained Classifier to Unlabeled End-User Data
Sleep apnea is a common yet severely under-diagnosed sleep related disorder. Unattended sleep monitoring at home with low-cost sensors can be leveraged for condition detection, and Machine Learning offers a generalized solution for this task. However, patient characteristics, lack of sufficient training data, and other factors can imply a domain shift between training and end-user data and reduced task performance. In this work, we address this issue with the aim to achieve personalization based on the patient’s needs. We present an unsupervised domain adaptation (UDA) solution with the constraint that labeled source data are not directly available. Instead, a classifier trained on the source data is provided. Our solution iteratively labels target data sub-regions based on classifier beliefs, and trains new classifiers from the expanding dataset. Experiments with sleep monitoring datasets and various sensors show that our solution outperforms the classifier trained on the source domain, with a kappa coefficient improvement from 0.012 to 0.242. Additionally, we apply our solution to digit classification DA between three well-established datasets, to investigate its generalizability, and allow for related work comparisons. Even without direct access to the source data, it outperforms several well-established UDA methods in these datasets.
- Open Access
- 1 July 2022
- Published by International Joint Conferences on Artificial Intelligence
- Vol. 6, 5762-5766
Learning Realistic Patterns from Visually Unrealistic Stimuli: Generalization and Data Anonymization (Extended Abstract)
Good training data is a prerequisite to develop useful Machine Learning applications. However, in many domains existing data sets cannot be shared due to privacy regulations (e.g., from medical studies). This work investigates a simple yet unconventional approach for anonymized data synthesis to enable third parties to benefit from such anonymized data. We explore the feasibility of learning implicitly from visually unrealistic, task-relevant stimuli, which are synthesized by exciting the neurons of a trained deep neural network. As such, neuronal excitation can be used to generate synthetic stimuli. The stimuli data is used to train new classification models. Furthermore, we extend this framework to inhibit representations that are associated with specific individuals. Extensive comparative empirical investigation shows that different algorithms trained on the stimuli are able to generalize successfully on the same task as the original model.
- 1 July 2022
- Published by International Joint Conferences on Artificial Intelligence
- Vol. 6, 5762-5766
Open AccessLearning Realistic Patterns from Visually Unrealistic Stimuli: Generalization and Data Anonymization (Extended Abstract)
Good training data is a prerequisite to develop useful Machine Learning applications. However, in many domains existing data sets cannot be shared due to privacy regulations (e.g., from medical studies). This work investigates a simple yet unconventional approach for anonymized data synthesis to enable third parties to benefit from such anonymized data. We explore the feasibility of learning implicitly from visually unrealistic, task-relevant stimuli, which are synthesized by exciting the neurons of a trained deep neural network. As such, neuronal excitation can be used to generate synthetic stimuli. The stimuli data is used to train new classification models. Furthermore, we extend this framework to inhibit representations that are associated with specific individuals. Extensive comparative empirical investigation shows that different algorithms trained on the stimuli are able to generalize successfully on the same task as the original model.
- Open Access
- 3 December 2021
- Published by AI Access Foundation in Journal of Artificial Intelligence Research
- Vol. 72, 1163-1214
Learning Realistic Patterns from Visually Unrealistic Stimuli: Generalization and Data Anonymization
Good training data is a prerequisite to develop useful Machine Learning applications. However, in many domains existing data sets cannot be shared due to privacy regulations (e.g., from medical studies). This work investigates a simple yet unconventional approach for anonymized data synthesis to enable third parties to benefit from such anonymized data. We explore the feasibility of learning implicitly from visually unrealistic, task-relevant stimuli, which are synthesized by exciting the neurons of a trained deep neural network. As such, neuronal excitation can be used to generate synthetic stimuli. The stimuli data is used to train new classification models. Furthermore, we extend this framework to inhibit representations that are associated with specific individuals. We use sleep monitoring data from both an open and a large closed clinical study, and Electroencephalogram sleep stage classification data, to evaluate whether (1) end-users can create and successfully use customized classification models, and (2) the identity of participants in the study is protected. Extensive comparative empirical investigation shows that different algorithms trained on the stimuli are able to generalize successfully on the same task as the original model. Architectural and algorithmic similarity between new and original models play an important role in performance. For similar architectures, the performance is close to that of using the original data (e.g., Accuracy difference of 0.56%-3.82%, Kappa coefficient difference of 0.02-0.08). Further experiments show that the stimuli can provide state-ofthe-art resilience against adversarial association and membership inference attacks.
- 3 December 2021
- Published by AI Access Foundation in Journal of Artificial Intelligence Research
- Vol. 72, 1163-1214
Open AccessLearning Realistic Patterns from Visually Unrealistic Stimuli: Generalization and Data Anonymization
Good training data is a prerequisite to develop useful Machine Learning applications. However, in many domains existing data sets cannot be shared due to privacy regulations (e.g., from medical studies). This work investigates a simple yet unconventional approach for anonymized data synthesis to enable third parties to benefit from such anonymized data. We explore the feasibility of learning implicitly from visually unrealistic, task-relevant stimuli, which are synthesized by exciting the neurons of a trained deep neural network. As such, neuronal excitation can be used to generate synthetic stimuli. The stimuli data is used to train new classification models. Furthermore, we extend this framework to inhibit representations that are associated with specific individuals. We use sleep monitoring data from both an open and a large closed clinical study, and Electroencephalogram sleep stage classification data, to evaluate whether (1) end-users can create and successfully use customized classification models, and (2) the identity of participants in the study is protected. Extensive comparative empirical investigation shows that different algorithms trained on the stimuli are able to generalize successfully on the same task as the original model. Architectural and algorithmic similarity between new and original models play an important role in performance. For similar architectures, the performance is close to that of using the original data (e.g., Accuracy difference of 0.56%-3.82%, Kappa coefficient difference of 0.02-0.08). Further experiments show that the stimuli can provide state-ofthe-art resilience against adversarial association and membership inference attacks.
- 1 February 2002
- Published by S. Karger AG in Neonatology
- Vol. 81 (2) , 119-125
Prolonged Apneas and Hypoxia Mediated by Nicotine and Endotoxin in Piglets
Objective: Infections and maternal smoking are risk factors for SIDS, and toxins from common bacteria have been proposed as a causative link between infections and SIDS. Nicotine can be transferred in significant amounts postnatally to the infant through environmental tobacco smoke or maternal smoking before nursing. We investigated the acute effects of nicotine and endotoxin on repeated apnea by laryngeal reflex stimulation and the following autoresuscitation. Design: Thirty-four 1-week-old (± 1 day) piglets were sedated and randomized to 1 of 4 pretreatment groups: (1) 1 µg endotoxin i.v./kg; (2) 5 µg nicotine i.v./kg; (3) 1 µg endotoxin i.v./kg and 5 µg nicotine i.v./kg, and (4) placebo. Apnea was induced by insufflation of 0.1 ml of acidified saline (pH 2) in the subglottic space of the trachea three times with intervals of 2 min. Results: Pretreatment with endotoxin caused a significant increase in plasma TNF-α in the endotoxin groups (mean value ± SEM 953 ± 246 and 980 ± 226 pg/ml, respectively) but no significant change in plasma IL-1β. Blood pressure, respiratory rate or SaO2 was not significantly affected before induced apnea. Both pretreatment with nicotine and endotoxin caused prolonged apneas by 35–45%, a complete loss of normal hyperventilation during autoresuscitation, and prolonged hypoxia after apnea. Conclusions: Nicotine and endotoxin interferes with autoresuscitation after apnea. This experimental model on piglets may shed light over important mechanisms involved in the causation of SIDS.
- 1 February 2002
- Published by S. Karger AG in Neonatology
- Vol. 81 (2) , 119-125
Prolonged Apneas and Hypoxia Mediated by Nicotine and Endotoxin in Piglets
Objective: Infections and maternal smoking are risk factors for SIDS, and toxins from common bacteria have been proposed as a causative link between infections and SIDS. Nicotine can be transferred in significant amounts postnatally to the infant through environmental tobacco smoke or maternal smoking before nursing. We investigated the acute effects of nicotine and endotoxin on repeated apnea by laryngeal reflex stimulation and the following autoresuscitation. Design: Thirty-four 1-week-old (± 1 day) piglets were sedated and randomized to 1 of 4 pretreatment groups: (1) 1 µg endotoxin i.v./kg; (2) 5 µg nicotine i.v./kg; (3) 1 µg endotoxin i.v./kg and 5 µg nicotine i.v./kg, and (4) placebo. Apnea was induced by insufflation of 0.1 ml of acidified saline (pH 2) in the subglottic space of the trachea three times with intervals of 2 min. Results: Pretreatment with endotoxin caused a significant increase in plasma TNF-α in the endotoxin groups (mean value ± SEM 953 ± 246 and 980 ± 226 pg/ml, respectively) but no significant change in plasma IL-1β. Blood pressure, respiratory rate or SaO2 was not significantly affected before induced apnea. Both pretreatment with nicotine and endotoxin caused prolonged apneas by 35–45%, a complete loss of normal hyperventilation during autoresuscitation, and prolonged hypoxia after apnea. Conclusions: Nicotine and endotoxin interferes with autoresuscitation after apnea. This experimental model on piglets may shed light over important mechanisms involved in the causation of SIDS.