Skip to content
This repository has been archived by the owner on Mar 14, 2023. It is now read-only.

Latest commit

 

History

History
50 lines (37 loc) · 3.82 KB

event.md

File metadata and controls

50 lines (37 loc) · 3.82 KB

Event

User population

Event- & trial-based experiments have an extensive history in behavioral and cognitive psychology. Fundamentally, data collection is carved up in time according to some ontology. Researchers may repeat Trial conditions in some manner to improve statistical power when contrasting a feature of interest versus a neutral baseline. Neuroscientists, in particular, may be interested in the moments before and after an Event to look at neurophysiological factors that predict or are predicted by a subject's behavior. What may differ between research groups is the ontology used to carve up time.

Key Projects

DataJoint has partnered with the following teams to interview key members, and develop individualized pipelines. By comparing across use-cases, the DataJoint team has developed a highly adaptable workflow to meet most needs, and trialize analyses within an existing DataJoint workflow.

  • International Brain Lab
  • Mesoscale Activity Project (Janelia Research Campus/Baylor College of Medicine/New York University)
  • Jerry Chen Lab (Boston University)
  • University of Rochester-New York University-Harvard University U19
  • Columbia University U19
  • Tobias Rose Lab (University of Bonn)

Pipeline Development

In addition to the key projects listed above, the DataJoint team met with leaders from both Neurodata Without Borders and the Kepecs Lab, as these groups have both tackled the difficulty of developing ontologies that can cover all possible iterations of behavioral data collection. Our resulting structure is exemplified by the figure below. The language below is tailored to the dependent variable in many neuroscience experiments, behavior.

|----------------------------------------------------------------------------|
|-------------------------------- Session ---------------------------------|__
|------------------------------- Recording ------------------------------|____
|----- Block 1 -----|______|----- Block 2 -----|______|----- Block 3 -----|___
| Trial 1 || Trial 2 |____| Trial 3 || Trial 4 |____| Trial 5 |____| Trial 6 |
|_|e1|_|e2||e3|_|e4|__|e5|__|e6||e7||e8||e9||e10||e11|____|e12||e13|_________|
|----------------------------------------------------------------------------|
  • A Session is period during which data is collected.
  • A Recording is some source of data tied to a single modality (e.g., behavior). This may or may not fully capture the session depending on recording latencies or equipment malfunctions.
  • Block and Trial are non-instantaneous subsets of Session whose traits often repeat across instances. These periods may be combined or contrasted in downstream analyses.
  • Trials may occur within or extend to the intervals between Blocks.
  • An Event (represented with e above) is an optionally instantaneous occurrence during a Session.
    • Projects may differ in their need to record event duration (e.g., onset versus duration of subject behavior)
    • Events may occur during other categories, or during continuously recorded behavior.

Features of Element Event include:

  • Pairing of upstream sessions with behavioral recordings
  • Multiple recorded attributes for phases of interest (see Attribute part tables for Block and Trial)
  • Defining Trial and Event Types as lookup tables
  • Optionally activating only the event schema for event-based recording, without Trial and Block phases.
  • An AlignmentEvent table to define the window of interest relative to specific event types.

Each level of the hierarchy (Block, Trial, Event) is designed to be optional to suit a given experiment's needs. For example usage, visit our Array Electrophysiology Workflow.