PPi: Pretraining Brain Signal Model for Patient-independent Seizure Detection
Authors: Zhizhang Yuan, Daoze Zhang, YANG YANG, Junru Chen, Yafeng Li
NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive experiments show that PPi outperforms the SOTA baselines on two public datasets and a real-world clinical dataset collected by us, which demonstrates the effectiveness and practicability of PPi. |
| Researcher Affiliation | Collaboration | Zhizhang Yuan* Zhejiang University zhizhangyuan@zju.edu.cn Daoze Zhang* Zhejiang University zhangdz@zju.edu.cn Yang Yang Zhejiang University yangya@zju.edu.cn Junru Chen Zhejiang University jrchen_cali@zju.edu.cn Yafeng Li Nuozhu Technology Co., Ltd. yafeng.li@neurox.cn |
| Pseudocode | Yes | Algorithm 1 Background Representation Calculation |
| Open Source Code | Yes | Our code is available at https://github.com/yzz673/PPi_public. |
| Open Datasets | Yes | The public datasets used in our paper, MAYO and FNUSA [32], are collected from two institutions: the Mayo Clinic (Rochester, Minnesota, United States of America) and St. Anne s University Hospital (Brno, Czech Republic), respectively. The datasets are publicly available to use under CC0 license and might be downloaded from https://springernature.figshare.com/collections/Multicenter_ intracranial_EEG_dataset_for_classification_of_graphoelements_and_ artifactual_signals/4681208. |
| Dataset Splits | Yes | We randomly choose 5 groups as the source domains (4 of which are used for training and 1 for validation) and the remaining group serves as the target domain. |
| Hardware Specification | Yes | All experiments run on a Linux system with 2 CPUs (AMD EPYC 7H12 64-Core Processor) and 4 GPUs (NVIDIA Ge Force RTX 3090). |
| Software Dependencies | No | The paper mentions running experiments on a 'Linux system' but does not specify any software dependencies with version numbers (e.g., Python, PyTorch, specific libraries). |
| Experiment Setup | Yes | The contexts half length h (i.e. the length of context above or context below) is an important hyperparameter in PPi. Thus we evaluate the performance on the clinical dataset under different contexts half length (shown in Fig. 5). |