This technology is a framework for efficiently storing, sending, processing, and receiving real-time data from commercial neurotech devices and wearables.
Current neural data pipelines are mostly developed in research environments, where the number of devices generating neural data is small, and there is no need for simultaneous data access among distributed users. Additionally, data format and transmission protocols are not efficiently designed for intensive distributed computing since time and resource constraints are not significant. Although there are commercial frameworks for high volume data, there are no current offerings specifically designed for real-time multidimensional neural data. Anticipating the aggravation of these problems on commercial neurotech devices, a new framework is needed to facilitate scalable data storage, standardization, search, and retrieval across devices.
This technology provides a scalable framework for efficient storing, processing, and transmission of neural data. It employs a protocol-buffer to optimize neural data storage on time-series databases. In addition, a service-based ecosystem enables a distributed data processing pipeline, based on remote procedure call, where multiple devices can send data to be stored and processed externally. These features are implemented cross-platform so it can be used in multiple devices with different software or hardware, and is language and platform neutral. Such a framework can be used for data collection from distributed large-scale neurotech devices to train machine learning models and provide real-time services.
This technology has been validated with a neurological dataset.
Patent Pending (US 20230344907)
IR CU22254, CU22255, CU23007, CU23292
Licensing Contact: Greg Maskel
