Smooth Curves, Similar Conclusions? Comparing Linear Regression and GAMLSS Neuropsychological Norms

Abstract

ObjectiveRegression-based normative approaches are widely used in neuropsychology but often rely on score transformations to satisfy model assumptions. We compared previously published linear regression (LR)-based norms with norms derived using Generalized Additive Models for Location, Scale and Shape (GAMLSS) for the brief cognitive battery used in the Norwegian Dementia Disease Initiation (DDI) cohort.

MethodGAMLSS norms were developed using the same normative samples as the original LR norms for the Consortium to Establish a Registry for Alzheimers Disease (CERAD) word list test, Trail Making Test (TMT) A and B, FAS phonemic fluency, and Visual Object and Space Perception Battery (VOSP) Silhouettes. Expected low-score frequencies and empirical base rates were assessed in a normative subsample (n = 131). Clinical implications were evaluated in the DDI clinical cohort (n = 643) using Mild Cognitive Impairment (MCI) classification, two-year diagnostic stability and change, and cerebrospinal fluid (CSF) biomarkers.

ResultsCompared with LR norms, GAMLSS yielded lower frequencies of low scores, primarily driven by CERAD delayed recall. Nevertheless, concordance between approaches was high ({kappa} = 0.91), with only 4.2% discordant classifications. Two-year diagnostic stability and change were broadly similar across approaches, and CSF biomarker profiles did not clearly favor either normative method.

ConclusionsGAMLSS provided a more faithful representation of neuropsychological score distributions, particularly for bounded and non-normal outcomes. However, downstream clinical differences were modest in this setting, suggesting that well-calibrated LR norms may remain robust for clinical classification.

Comparing the Phenotypic and Genetic Differences Between Autistic Regression and Autism Without Regression Using the SPARK Cohort
A. Yahia, Y. Zhang, A. Gkerdi +2
medRxiv26 Oct 2026PreprintCC BY-NDFull textHuman geneticsmedRxiv
Cognitive phenotypes and neuropsychological trajectories in post-COVID-19 syndrome: a 12-month follow-up study
Elena Herrera, Ainhoa Señoráns, Arrate Barrenechea +2
Scientific Reports5 Oct 2026Peer-reviewedCC BY-NC-ND 4.0Full textNeurologyCrossref
Post-Stroke Cognitive Impairment in Georgia: Diagnostic Utility of Georgian NINDS-CSN Neuropsychological Battery
Maia Tighashvili, Nina Mikeladze, Nazibrola Botchorishvili +5
Neurology international23 Sep 2026Peer-reviewedCC BYFull textGeneral medicineEurope PMC
International variation in neuropsychological test use practices: Implications for test adaptation and fair assessment
Porrselvi A.P, Lena Dobson, Brandon Frank +11
Applied neuropsychology. Adult18 Sep 2026Peer-reviewedLicence not statedFull textPsychiatry & mental healthPMC
Data Extraction Errors and Diagnostic Conclusions of Meta-Analyses
Zixing Huang, Yuan Tian, Ping Gong +2
JAMA Network Open1 Oct 2026Peer-reviewedCC BY-NC-NDFull textGeneral medicineCrossrefEurope PMCPMC

Cite

BibTeX RIS

BibTeX

@misc{kirsebom2026smooth,
  title = {{Smooth Curves, Similar Conclusions? Comparing Linear Regression and GAMLSS Neuropsychological Norms}},
  author = {B.-E. Kirsebom and I. Myrvoll Lorentzen and J. Espenes and I. Vollo Eliassen and F. Gonzalez-Ortiz and A. Wallin and K. Waterloo and M. Eckerstrom and G. Rolfseng Grontvedt and E. Hessen and T. Fladby},
  howpublished = {medRxiv},
  year = {2026},
  month = sep,
  doi = {10.64898/2026.09.01.26361590},
  url = {https://www.medrxiv.org/content/10.64898/2026.09.01.26361590v1},
  note = {CC BY-NC-ND},
}

RIS

TY  - UNPB
TI  - Smooth Curves, Similar Conclusions? Comparing Linear Regression and GAMLSS Neuropsychological Norms
AU  - B.-E. Kirsebom
AU  - I. Myrvoll Lorentzen
AU  - J. Espenes
AU  - I. Vollo Eliassen
AU  - F. Gonzalez-Ortiz
AU  - A. Wallin
AU  - K. Waterloo
AU  - M. Eckerstrom
AU  - G. Rolfseng Grontvedt
AU  - E. Hessen
AU  - T. Fladby
PY  - 2026
DA  - 2026/09/01
PB  - medRxiv
DO  - 10.64898/2026.09.01.26361590
UR  - https://www.medrxiv.org/content/10.64898/2026.09.01.26361590v1
AB  - ObjectiveRegression-based normative approaches are widely used in neuropsychology but often rely on score transformations to satisfy model assumptions. We compared previously published linear regression (LR)-based norms with norms derived using Generalized Additive Models for Location, Scale and Shape (GAMLSS) for the brief cognitive battery used in the Norwegian Dementia Disease Initiation (DDI) cohort.  MethodGAMLSS norms were developed using the same normative samples as the original LR norms for the Consortium to Establish a Registry for Alzheimers Disease (CERAD) word list test, Trail Making Test (TMT) A and B, FAS phonemic fluency, and Visual Object and Space Perception Battery (VOSP) Silhouettes. Expected low-score frequencies and empirical base rates were assessed in a normative subsample (n = 131). Clinical implications were evaluated in the DDI clinical cohort (n = 643) using Mild Cognitive Impairment (MCI) classification, two-year diagnostic stability and change, and cerebrospinal fluid (CSF) biomarkers.  ResultsCompared with LR norms, GAMLSS yielded lower frequencies of low scores, primarily driven by CERAD delayed recall. Nevertheless, concordance between approaches was high ({kappa} = 0.91), with only 4.2% discordant classifications. Two-year diagnostic stability and change were broadly similar across approaches, and CSF biomarker profiles did not clearly favor either normative method.  ConclusionsGAMLSS provided a more faithful representation of neuropsychological score distributions, particularly for bounded and non-normal outcomes. However, downstream clinical differences were modest in this setting, suggesting that well-calibrated LR norms may remain robust for clinical classification.
ER  -

Abstract

Related

Cite