Neo

Neo is a Python package designed for working with electrophysiology data. It provides a common, shared object model to improve interoperability between various Python tools used for analyzing, visualizing, and generating neurophysiology data.

Key Features:

• Broad Format Support: Reads a wide range of neurophysiology formats including Spike2, NeuroExplorer, AlphaOmega, Axon, Blackrock, Plexon, Tdt, and Igor Pro.
• Data Writing: Supports writing to a subset of the above formats, as well as non-proprietary formats like Kwik and HDF5.
• Hierarchical Data Model: Adapted for intracellular, extracellular, and EEG data, including support for multi-electrode setups (e.g., tetrodes).
• NumPy Integration: Neo objects are built on the quantities package and NumPy. They behave like normal NumPy arrays but include metadata, dimensional consistency checks, and automatic unit conversion.

Neo is a Python package designed for working with electrophysiology data. It provides a common, shared object model to improve interoperability between different Python tools used for analyzing, visualizing, and generating electrophysiology data.

Key Features:

• Broad Format Support: Reads a wide range of neurophysiology formats including Spike2, NeuroExplorer, AlphaOmega, Axon, Blackrock, Plexon, and Tdt. It can write to a subset of these and non-proprietary formats like HDF5.
• Hierarchical Data Model: Adapted for intracellular, extracellular, and EEG data, with support for multi-electrode configurations (e.g., tetrodes).
• NumPy-Compatible Objects: Neo objects are built on the quantities package and NumPy. They behave like standard NumPy arrays but include metadata, dimensional consistency checks, and automatic unit conversion.
• Lightweight Design: Neo focuses strictly on data representation and does not include built-in functions for data analysis or visualization.

Neo is a Python package designed for working with electrophysiology data, including intracellular, extracellular, and EEG data. It provides a hierarchical data model aimed at improving interoperability between different Python tools for analyzing, visualizing, and generating neurophysiology data by offering a common, shared object model.

Key features include:

• Wide Format Support: Reading many neurophysiology formats (e.g., Spike2, NeuroExplorer, AlphaOmega, Axon, Blackrock, Plexon, Tdt, Igor Pro) and writing to open formats like NWB and NIX.
• NumPy-like Behavior: Neo objects behave like standard NumPy arrays but include additional metadata, dimensional consistency checks, and automatic unit conversion.
• Lightweight Design: Neo focuses strictly on the representation of data; it does not include built-in functions for data analysis or visualization to keep dependencies minimal.

Neo data objects support three levels of metadata:

1. Required Metadata: Attributes that must exist for the object to be valid. For an AnalogSignal, this includes sampling_rate.
2. Recommended/Optional Metadata: Attributes like name that appear directly on the object.
3. Fully Optional Metadata: Stored in two specific attributes:
   • annotations: A dictionary for general metadata.
   • array_annotations: A dictionary for metadata that is array-like (e.g., mapping channel indices to specific values). Slicing the data object automatically slices these annotations.

When slicing a Neo object, the array_annotations are updated to match the slice.

# Accessing metadata
print(signal.name)
print(signal.sampling_rate)
print(signal.annotations)
print(signal.array_annotations)

# Slicing preserves array_annotations
sliced_signal = signal[100:110, 1:3]
print(sliced_signal.array_annotations)

When a Block contains SpikeTrain objects, you can access them using three different organizational strategies depending on your analysis goal:

• By Trial (Segment): Access all spike trains within a specific segment to calculate metrics like a Peristimulus Time Histogram (PSTH) for a single trial.
• By Neuron (Group): Access spike trains within a specific group to analyze the activity of an individual unit across all trials.
• By Tetrode (Filtering): Use the block.filter() method to find groups associated with a specific hardware identifier (e.g., a tetrode_id stored in annotations) to blend activity from all units on that tetrode.

# By trial: PSTH for each segment
for seg in block.segments:
    stlist = [st - st.t_start for st in seg.spiketrains]
    # ... calculate histogram ...

# By neuron: PSTH for each group (unit)
for group in block.groups:
    stlist = [st - st.t_start for st in group.spiketrains]
    # ... calculate histogram ...

# By tetrode: Blend all units on a specific tetrode
for tetrode_id in block.annotations["tetrode_ids"]:
    stlist = []
    for unit in block.filter(objects=Group, tetrode_id=tetrode_id):
        stlist.extend([st - st.t_start for st in unit.spiketrains])
    # ... calculate histogram ...