diff --git a/doc/sam.qmd b/doc/sam.qmd
new file mode 100644
index 0000000..63074a8
--- /dev/null
+++ b/doc/sam.qmd
@@ -0,0 +1,111 @@
+---
+title: SAM
+format:
+  html:
+    toc: true
+    toc-title: Contents
+    code-block-bg: true
+    code-block-border-left: "#31BAE9"
+    code-line-numbers: true
+    highlight-style: atom-one
+    link-external-icon: true
+    link-external-newwindow: true
+    eqn-number: true
+---
+
+### SAM
+
+```{python}
+import pathlib
+
+import numpy as np
+import plotly.graph_objects as go
+import rlxnix as rlx
+import scipy.signal as signal
+from plotly.subplots import make_subplots
+from util import plot_line_comparision, trial_plot
+
+dataset_path = pathlib.Path("../oephys2nix/test/AllStimuli/2025-10-20-aa-invivo-2-recording.nix")
+relacs_path = pathlib.Path(
+    "../oephys2nix/test/AllStimuli/2025-10-20-aa-invivo-2_relacs/2025-10-20-aa-invivo-2_relacs.nix"
+)
+
+dataset = rlx.Dataset(str(dataset_path))
+relacs = rlx.Dataset(str(relacs_path))
+
+# INFO: Select the first stimulus of the calibration repro
+repro_d = dataset.repro_runs("SAM")[0].stimuli[2]
+repro_r = relacs.repro_runs("SAM")[0].stimuli[2]
+
+sinus, t = repro_d.trace_data("sinus")
+sinus_r, t_r = repro_r.trace_data("V-1")
+
+stimulus_oe, t = repro_d.trace_data("stimulus")
+stimulus_re, t_r = repro_r.trace_data("GlobalEFieldStimulus")
+
+local_eod_oe, t = repro_d.trace_data("local-eod")
+local_eod_re, t_r = repro_r.trace_data("LocalEOD-1")
+
+global_eod_oe, t = repro_d.trace_data("global-eod")
+global_eod_re, t_r = repro_r.trace_data("EOD")
+
+ttl, t = repro_d.trace_data("ttl-line")
+```
+### Plotting the First trial
+If you zoom in you can see a little delay between the different recording systems. It seems that open-ephys is before the relacs recording.
+
+```{python}
+# | echo: False
+fig = trial_plot(repro_d, repro_r, 0.2)
+fig.update_xaxes(range=[0, 0.2])
+fig.show()
+```
+### Correlation between the Signals
+
+```{python}
+print(f"Duration of the dataset {repro_d.duration}")
+print(f"Duration of the relacs {repro_r.duration}")
+# Resample the open-ephys data
+sinus_resampled = signal.resample(sinus, len(sinus_r))
+```
+
+```{python}
+# | echo: False
+fig = plot_line_comparision(
+    t_r, t_r, sinus_r, sinus_resampled, ["sinus-relacs", "sinus-resampled-open-ephys"]
+)
+fig.show()
+```
+We need to scale the two signals
+
+```{python}
+oephys_lanes = [sinus, local_eod_oe, global_eod_oe, stimulus_oe]
+relacs_lanes = [sinus_r, local_eod_re, global_eod_re, stimulus_re]
+names_lanes = ["sinus", "local-eod", "global-eod", "stimulus"]
+lags_lanes = []
+for oephys_lane, relacs_lane, names_lane in zip(
+    oephys_lanes, relacs_lanes, names_lanes, strict=True
+):
+    print(oephys_lane.shape)
+    print(relacs_lane.shape)
+    oephys_lane_resampled = signal.resample(oephys_lane, len(relacs_lane))
+    correlation = signal.correlate(oephys_lane_resampled, relacs_lane, mode="full")
+    lags = signal.correlation_lags(oephys_lane_resampled.size, relacs_lane.size, mode="full")
+    lag = lags[np.argmax(correlation)]
+    lags_lanes.append(lag)
+    print(f"{names_lane} has a lag of {lag}")
+```
+
+```{python}
+# | echo: False
+fig = plot_line_comparision(
+    t_r,
+    t,
+    np.roll(stimulus_re, lags_lanes[-1]),
+    stimulus_oe - np.mean(stimulus_oe),
+    ["rolled sinus-relacs", "sinus-resampled-open-ephys"],
+)
+fig.show()
+
+print(f"The lag of the whitenoise is {lags_lanes[-1] * (1 / 20_000) * 1000} milli seconds")
+```