Serial Repetition Lag-CRP

Analyze repetition contiguity conditioned on correctly ordered recall.

The serial repetition CRP conditions on correctly ordered output up to the point of analysis. It tabulates repetition-index CRPs only for transitions where the recall sequence matches study order, stopping once an out-of-order response occurs. This isolates temporal contiguity effects in participants who are maintaining serial order.

Workflow

Code

import os

import jax.numpy as jnp
import matplotlib.pyplot as plt

from jaxcmr.analyses.serialrepcrp import plot_rep_crp, subject_serial_rep_crp, test_serial_rep_crp_vs_control
from jaxcmr.helpers import find_project_root, generate_trial_mask, load_data, save_figure
from jaxcmr.repetition import make_control_dataset

Code

allow_repeated_recalls = True
filter_repeated_recalls = False
data_path = "data/RepeatedRecallsKahanaJacobs2000.h5"
figure_dir = "results/figures"
figure_str = ""
ylim = [0, .9]
base_data_name = "KahanaJacobs2000"
# trial_query = "data['condition'] == 2"
# mixed_trial_query = "jnp.logical_and(data['condition'] == 2, data['lag'] != 4)"
# control_trial_query = "jnp.logical_and(data['condition'] == 2, data['lag'] == 4)"
trial_query = "jnp.logical_and(data['recall_attempt'] == 1, data['recall_total'] > 0)"

mixed_trial_query = "jnp.logical_and(jnp.logical_and(data['recall_attempt'] == 1, data['recall_total'] > 0), data['repetitions'] == 1)"
control_trial_query = "jnp.logical_and(jnp.logical_and(data['recall_attempt'] == 1, data['recall_total'] > 0), data['repetitions'] == 0)"
control_shuffles = 10
control_seed = 0
min_lag = 2
max_lag = 3
size = 2

Code

project_root = find_project_root()
figure_dir = os.path.join(project_root, figure_dir)
data_path = os.path.join(project_root, data_path)
data = load_data(data_path)
trial_mask = generate_trial_mask(data, trial_query)
mixed_mask = generate_trial_mask(data, mixed_trial_query) & trial_mask

control_dataset = make_control_dataset(
    data,
    mixed_query=mixed_trial_query,
    control_query=control_trial_query,
    n_shuffles=control_shuffles,
    remove_repeats=filter_repeated_recalls,
    seed=control_seed,
)
control_mask = jnp.ones((control_dataset["recalls"].shape[0],), dtype=bool)

Code

plot_rep_crp(
    datasets=data,
    trial_masks=mixed_mask,
    max_lag=max_lag,
    min_lag=min_lag,
    size=size,
    labels=("First", "Second"),
    contrast_name="Repetition Index",
)
plt.title("Mixed Trials")
if ylim is not None:
    for ax in plt.gcf().axes:
        ax.set_ylim(ylim)
save_figure(figure_dir, figure_str, suffix="mixed")

plot_rep_crp(
    datasets=control_dataset,
    trial_masks=control_mask,
    max_lag=max_lag,
    min_lag=min_lag,
    size=size,
    labels=("First", "Second"),
    contrast_name="Repetition Index",
)
plt.title("Control (Shuffled) Trials")
if ylim is not None:
    for ax in plt.gcf().axes:
        ax.set_ylim(ylim)
save_figure(figure_dir, figure_str, suffix="control")

Code

observed_crp = subject_serial_rep_crp(data, mixed_mask, min_lag, max_lag, size)
control_crp = subject_serial_rep_crp(control_dataset, control_mask, min_lag, max_lag, size)

results = test_serial_rep_crp_vs_control(observed_crp, control_crp, max_lag)
for label, result in results.items():
    print()
    print("=" * 60)
    print(f"Statistical Tests: {label}")
    print("=" * 60)
    print(result)


============================================================
Statistical Tests: First Presentation
============================================================
  Lag |   t-stat    t p-val |   W-stat    W p-val |  Mean Diff
------+----------------------+----------------------+------------
    1 |   -1.648     0.1167 |     54.0     0.1042 |    -0.0158
    2 |    0.999     0.3309 |     68.0     0.2935 |     0.0042
    3 |    1.192     0.2486 |     66.0     0.2579 |     0.0061

============================================================
Statistical Tests: Second Presentation
============================================================
  Lag |   t-stat    t p-val |   W-stat    W p-val |  Mean Diff
------+----------------------+----------------------+------------
    1 |    1.814     0.0864 |     54.0     0.1042 |     0.0076
    2 |   -0.511     0.6154 |     81.0     0.5949 |    -0.0013
    3 |   -1.390     0.1815 |     25.0     0.0843 |    -0.0015

Interpretation

Plots show repetition CRP curves computed only from correctly ordered recall prefixes. Key patterns:

Stronger contiguity under order: if serial-order maintenance amplifies temporal contiguity, curves will be steeper than unconditioned repetition CRP.
Observed vs. control: differences isolate repetition-specific effects in order-maintaining recall.
First vs. second presentation: reveals which occurrence drives transitions when participants recall in order.

API Details

Notebook parameters

This template uses the same parameter set as the repetition Lag-CRP template. See that template’s API Details for parameter descriptions. Additional parameters:

allow_repeated_recalls — whether the analysis permits repeated recalls.
filter_repeated_recalls — whether to filter out repeated recall events.