Error when using fitrm - Fit Repeated Measures Model for RM ANOVA

Neeta Dsouza answered . 2025-03-27 15:44:18

The error you're encountering typically occurs when there are missing (NaN) or infinite (Inf) values in the input data matrix, which prevents MATLAB from calculating eigenvalues as part of the statistical analysis.

Here are some steps to help you troubleshoot and resolve the issue:

1. Check Your Data for NaN or Inf Values

Before using the fitrm function, ensure your data does not contain missing or infinite values. You can check your data table as follows:

% Check for NaN or Inf in your data
any(isnan(table2array(yourTable)), 'all')  % Replace 'yourTable' with your data table
any(isinf(table2array(yourTable)), 'all')

If these checks return true, it means your data contains problematic values. You can either remove or impute these values, depending on the situation.

2. Verify Data Table Format

Ensure your data is formatted correctly for fitrm. Typically, the data should be in a table where:

• Each row represents a participant.
• Columns include the repeated measures and any between-subject variables.

For example:

data = table(...
    {'P1'; 'P2'; 'P3'; 'P4'}, ...
    [1.2, 1.3, 1.4, 1.5]', ...
    [2.1, 2.3, 2.5, 2.6]', ...
    [3.4, 3.6, 3.8, 4.0]', ...
    [4.5, 4.7, 4.8, 5.0]', ...
    'VariableNames', {'Participant', 'L1', 'L2', 'L3', 'L4'});

3. Define the Model Correctly

Make sure your formula for fitrm matches your data. For example:

rm = fitrm(data, 'L1-L4 ~ 1', 'WithinDesign', table([1; 2; 3; 4], 'VariableNames', {'Location'}));

4. Debugging Example

Here's a minimal example that demonstrates proper formatting:

% Create example data
data = table(...
    {'P1'; 'P2'; 'P3'; 'P4'}, ...
    [1.2; 1.3; 1.4; 1.5], ...
    [2.1; 2.3; 2.5; 2.6], ...
    [3.4; 3.6; 3.8; 4.0], ...
    [4.5; 4.7; 4.8; 5.0], ...
    'VariableNames', {'Participant', 'L1', 'L2', 'L3', 'L4'});

% Define within-subjects factors
within = table({'L1'; 'L2'; 'L3'; 'L4'}, 'VariableNames', {'Location'});

% Fit repeated measures model
rm = fitrm(data, 'L1-L4 ~ 1', 'WithinDesign', within);

% Display results
ranova(rm)

5. Inspect the Output

After running the above example, verify that no errors occur. If this works but your original data fails, you can iteratively compare the two to identify the issue.

6. Additional Notes

• If your actual data has a much longer list of measurements, ensure the column names follow MATLAB's naming conventions (e.g., no special characters).
• If you need to handle missing data, consider replacing NaN values with group means or other imputation methods before fitting the model.

If you're still encountering issues, feel free to share more details or your dataset (an anonymized sample), and I can assist further!

Not satisfied with the answer ?? ASK NOW