### Install Dependencies and Build Cython Wrapper

Source: https://github.com/mtegu/aep-model/blob/master/README.md

This snippet shows the commands to set up a virtual environment, install project dependencies, and build the Cython wrapper for the AN model. Ensure you are in the correct directory before running.

```bash
python -m venv venv

# macOS/Linux
source venv/bin/activate

# Windows (PowerShell)
.\venv\Scripts\Activate.ps1

pip install -r requirements.txt

cd BEZ2018model
python setup.py build_ext --inplace
```

--------------------------------

### Run Basic Auditory Nerve Simulation

Source: https://github.com/mtegu/aep-model/blob/master/README.md

Execute a basic simulation of the auditory nerve model after installation. The output, containing simulated auditory nerve activity per CF, will be saved as 'results.pkl'.

```bash
cd ..
python test_run_ANmodel.py
```

--------------------------------

### Initialize Simulated CAP Arrays

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Initializes arrays to store simulated click-evoked CAPs and sets up the corresponding time vector. The simulated CAPs will be generated by convolving auditory nerve responses with the derived unitary responses.

```python
# Initialize the simulated click-evoked CAPs
simul_CAP = np.zeros((8, 20, an_click.shape[2] + ur.shape[1] - 1))
t_simul = np.arange(0, simul_CAP.shape[2])/output_fs*1000 - 10 - offset

print(f"Simulated CAP array shape: {simul_CAP.shape}")
print(f"Number of stimulus levels: {simul_CAP.shape[0]}")
print(f"Number of subjects: {simul_CAP.shape[1]}")
print(f"Simulated CAP length: {simul_CAP.shape[2]} samples")
print(f"Simulated time vector range: {t_simul[0]:.2f} to {t_simul[-1]:.2f} ms")
```

--------------------------------

### Prepare Data for Unitary Response Derivation

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Prepares the auditory nerve model output for unitary response derivation by setting up time vectors, defining offset parameters, and selecting an appropriate time window. Sums responses across characteristic frequencies.

```python
# Prepare the AN model output
offset = 1  # ms - This allows for a smoother start of the UR
t_vect = np.arange(0, an_click.shape[2])/output_fs*1000 - 10  # 10 ms silence pre stimulus onset
t_X_0 = np.argmin(np.abs(t_vect - (-10+offset)))  # Experimental CAPs start at -10 ms
t_X_10 = np.argmin(np.abs(t_vect - (30+offset)))  # Experimental CAPs end at 30 ms
X = np.sum(an_click[:,:, t_X_0:t_X_10], axis=1)  # add across all CFs

print(f"Time vector range: {t_vect[0]:.2f} to {t_vect[-1]:.2f} ms")
print(f"Selected time window: {t_vect[t_X_0]:.2f} to {t_vect[t_X_10-1]:.2f} ms")
print(f"Prepared data shape for UR derivation: {X.shape}")
print(f"Offset parameter: {offset} ms")
```

--------------------------------

### Import Libraries for Auditory Nerve Modeling

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Imports necessary libraries including NumPy, pickle, matplotlib, and custom utility functions for auditory nerve modeling and visualization. Sets the matplotlib backend for Windows compatibility.

```python
import numpy as np
import pickle
import matplotlib as mpl
if 'Qt5Agg' in mpl.rcsetup.all_backends: # for Windows
    mpl.use('Qt5Agg')
from utils import plot_tens_output, plot_group_average_CAPs
from simulator import derive_unitary_response
import matplotlib.pyplot as plt
```

--------------------------------

### Load Experimental Evoked Data

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Loads experimental evoked data from a .npz file, including the sampling rate, click-evoked CAP waveform, and time vector. Prints the details of the loaded experimental data.

```python
# Load experimental data
data = np.load("experimental_dataset/evoked_data.npz")
fs = data["fs"]  # Important that this matches the output_fs
experimental_cap = data["waveform_click_cap"]  # shape (20, n_time)
t_exp_data = data["time_vector"]

print(f"Experimental data sampling rate: {fs} Hz")
print(f"Experimental CAP waveform shape: {experimental_cap.shape}")
print(f"Time vector length: {len(t_exp_data)}")
print(f"Time range: {t_exp_data[0]:.2f} to {t_exp_data[-1]:.2f} ms")
```

--------------------------------

### Launch Jupyter Notebook for Unitary Response Derivation

Source: https://github.com/mtegu/aep-model/blob/master/README.md

This command launches the Jupyter Notebook used for deriving unitary responses (UR) by convolving AN fiber activity with a unitary response. This process is used to generate AEPs.

```bash
jupyter notebook test_derive_UR.ipynb
```

```bash
jupyter lab test_derive_UR.ipynb
```

--------------------------------

### Load Simulated Auditory Nerve Data

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Loads simulated auditory nerve data from a pickle file. Extracts averaged responses, characteristic frequency vector, and sampling rate. Prints the shapes and parameters of the loaded data.

```python
with open('simulation_results/results_clicks.pkl', 'rb') as f:
    results = pickle.load(f)
    
an_click = results['tens_dense_ave']
cf_vect = results['cf_vect']
output_fs = 10e3

print(f"Loaded auditory nerve data with shape: {an_click.shape} levels x channels x time points")
print(f"Characteristic frequency vector length: {len(cf_vect)}")
print(f"Output sampling rate: {output_fs} Hz")
```

--------------------------------

### Visualize Derived Unitary Response

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Creates a detailed plot of the grand average unitary response, including axis labels and formatting. This visualization illustrates the final derived unitary response characterizing the neural-to-electrical transformation.

```python
# Plot the derived unitary response
fig, ax = plt.subplots(figsize=(5, 5))
t_UR = np.arange(0, ur.shape[1])/output_fs*1000 - offset

ax.plot(t_UR, ur.mean(axis=0), color='k', linewidth=2)

ax.set_xlabel('Time (ms)')
ax.set_xlim([-offset, 9])
ax.set_ylim([-4e-4, 2e-4])
ax.set_ylabel('(a.u.)', y=0.25, x=1)
ax.set_title('Grand Average Unitary Response')
ax.set_yticks([ax.get_ylim()[0], ax.get_ylim()[0] + 5e-4])
ax.set_yticklabels(['0', '0.5e-4'])
ax.grid(True, alpha=0.3)

fig.tight_layout()
plt.show()
```

--------------------------------

### Visualize Auditory Nerve Model Output

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Visualizes the auditory nerve model output using the plot_tens_output function. Displays the response patterns across different characteristic frequencies and intensity levels.

```python
# Plot the AN model output by level
plot_tens_output(an_click, cf_vect, output_fs, title='AN Output', xlim=(0, 40), xlim2=(0, .1), ylim=(0, 1.5e3))
```

--------------------------------

### Derive Unitary Response

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Derives the unitary response using the derive_unitary_response function. This function optimizes to find the best unitary response that, when convolved with the AN model output, matches the experimental CAP data. Parameters for regularization are set.

```python
# Derive the unitary response
# Parameters: max_level=7, lambda=5 (minimum MSE regularization parameter)
ur = derive_unitary_response(7, X[:,:experimental_cap.shape[2]], experimental_cap, 5, plot=False)

print(f"Derived unitary response shape: {ur.shape}")
print(f"Number of subjects: {ur.shape[0]}")
print(f"Unitary response length: {ur.shape[1]} samples")
print(f"Unitary response duration: {ur.shape[1]/output_fs*1000:.2f} ms")
```

--------------------------------

### Convolve Auditory Nerve Responses with Unitary Responses

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Convolves auditory nerve responses with derived unitary responses for each stimulus level and subject to generate simulated CAPs. This is used to compare simulated data with experimental results.

```python
# Convolve click AN responses with URs for each level and subject
for lvl_idx in range(8):
    for sub_idx in range(20):
        signal = an_click[lvl_idx, :, :].sum(axis=0)  # add across all CFs
        simul_CAP[lvl_idx, sub_idx, :] = np.convolve(signal, ur[sub_idx, :])

print("Convolution completed for all stimulus levels and subjects")
print(f"Generated simulated CAPs with shape: {simul_CAP.shape}")
```

--------------------------------

### Plot Group Average CAPs

Source: https://github.com/mtegu/aep-model/blob/master/test_derive_UR.ipynb

Visualizes and compares experimental and simulated group average CAPs using the plot_group_average_CAPs function. This comparison helps validate the quality of the derived unitary response.

```python
# Plot the experimental and simulated CAPs
plot_group_average_CAPs(experimental_cap, t_exp_data, simul_CAP, t_simul, scale=0.6, title='Group Average Click CAPs')
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.