> ## Documentation Index
> Fetch the complete documentation index at: https://docs.autocleaneeg.org/llms.txt
> Use this file to discover all available pages before exploring further.

# Flexibility of the Task Class

> See how the Task class enables branching, custom functions, and integration with Python or MATLAB to adapt pipelines for any experimental design.

## Flexibility of the Task Class

The real strength of the Task class is its **flexibility**.\
Because each step works directly on an **MNE object** (`Raw`, `Epochs`, or `Evoked`), you can:

* Insert **custom functions** for specialized preprocessing.
* Branch into **different processing paths** depending on your experiment.
* Call **external libraries** (even MATLAB functions via `matlab.engine`) when needed.
* Return to the main pipeline as long as you hand back a valid MNE object.

### Example: Branching for Custom Epoching

```python theme={null}
class Oddball_Task(Task):
    def run(self):
        self.import_raw()
        self.filter_data()

        self.epochs = self.custom_event_epochs()
        
        self.generate_reports()

    def custom_event_epochs(self):
        """Epoch by event codes (e.g., standard vs oddball)."""
        events = mne.find_events(self.raw, stim_channel="STI 014")
        event_id = {"standard": 1, "oddball": 2}
        epochs = mne.Epochs(
            self.raw, events, event_id,
            tmin=-0.2, tmax=0.8, baseline=(None, 0), preload=True
        )
        return epochs
```

### Why This Matters

* **Customizable**: Tailor preprocessing to your study (e.g., resting state, oddball, Go/No-Go).
* **Branchable**: Choose different paths for different datasets in the same task.
* **Extensible**: Insert ML models, advanced artifact rejection, or external tool calls.
* **Future-Proof**: As long as the function returns an MNE object, the pipeline can continue without breaking.

<Info>
  Think of the Task class as the **backbone of your preprocessing recipe**.\
  You can keep it simple or expand it into a fully customized pipeline — the framework is designed to handle both.
</Info>