Data for: Daume et al. (2024) "Persistent activity during working memory maintenance predicts long-term memory formation in the human hippocampus"

This dataset contains the source data for Daume et al. (2024) "Persistent activity during working memory maintenance predicts long-term memory formation in the human hippocampus". Example code to analyze the data is deposited at Github (https://github.com/rutishauserlab/SBCAT-NO-release-NWB) and Zenodo (https://doi.org/10.5281/zenodo.13621888).

Abstract:
Working Memory (WM) and Long-Term Memory (LTM) are often viewed as separate cognitive systems. Little is known about how these systems interact when forming memories. We recorded single neurons in the human medial temporal lobe while patients maintained novel items in WM and a subsequent recognition memory test for the same items. In the hippocampus but not the amygdala, the level of WM content-selective persist activity during WM maintenance was predictive of whether the item was later recognized with high confidence or forgotten. In contrast, visually evoked activity in the same cells was not predictive of LTM formation. During LTM retrieval, memory-selective neurons responded more strongly to familiar stimuli for which persistent activity was high while they were maintained in WM. Our study suggests that hippocampal persistent activity of the same cell supports both WM maintenance and LTM encoding, thereby revealing a common single-neuron component of these two memory systems.

Note:
sub-35_ses-1_ecephys+image.nwb is missing waveform mean/std data, but they can be derived from the raw spike waveforms included in the file.