When a Simple Problem Stops Good Science

This study began with a deceptively simple problem. During behavioural testing at the research centre in Glostrup, rats placed in the light/dark box rarely entered the illuminated compartment at all. While this behaviour is unsurprising from the animal’s perspective, it posed a major challenge scientifically: if control animals avoid light almost completely, it becomes extremely difficult to detect treatment‑induced changes in light aversion. Before we could meaningfully study photophobia‑like behaviour, we needed a model that separated true light‑driven avoidance from anxiety and novelty stress.

Small Changes, Big Impact: Refining the Model and Rethinking CGRP Delivery

Our initial focus therefore became refinement of the model itself rather than the pharmacology. We explored several ideas, including food‑based motivation and shelters in the bright compartment, but the most effective solution turned out to be the simplest. By adding bedding material from the animals’ own home cages to both compartments of the light/dark box, we markedly reduced novelty‑induced anxiety. The effect was striking: even at high light intensities, naïve rats now spent close to equal time in light and dark. This refinement improved both animal welfare and experimental sensitivity, and notably, it is easy for other laboratories to implement.

With an optimised behavioural setup in place, we turned to addressing another gap in the preclinical migraine literature. While intravenous CGRP is widely used in human provocation studies, preclinical models almost exclusively rely on intraperitoneal or intracranial administration—routes that are either poorly controlled or highly invasive. We could not identify previous rodent studies investigating i.v. CGRP‑induced light aversion, despite the clear translational relevance. We therefore chose to administer CGRP intravenously to awake rats, using a single bolus dose to avoid anaesthesia and additional pharmacological confounders. This allowed us to evaluate CGRP’s effects under conditions that more closely resemble clinical provocation paradigms.

When Results Do Not Match Expectations: The Hidden Role of Housing and Stress

One of the most informative aspects of the study emerged unexpectedly. Based on existing literature, we anticipated that female rats would show equal or greater sensitivity to CGRP than males. Instead, in the first phase of the study, CGRP-injected females did not display the expected light‑avoidant behaviour, but rather appeared to spend slightly more time in the light compartment than the controls. This discrepancy prompted us to reconsider aspects of the experimental design that are often taken for granted, particularly housing conditions. In the initial setup, CGRP‑treated and control animals were co‑housed. When we later housed animals according to treatment group, the CGRP‑induced light aversion became evident in females as well.

This finding underscored the importance of social buffering and emotional contagion in behavioural studies, particularly in female rodents. It also reinforced one of the central conclusions of our work: factors such as housing, handling, and test order are not minor methodological details, but can fundamentally shape behavioural outcomes and even mask true pharmacological effects.

Alongside behavioural testing, we assessed cephalic mechanical sensitivity to determine whether light aversion occurred together with headache‑like pain. When tested in naïve animals housed by treatment group, intravenous CGRP induced clear periorbital allodynia in both sexes, while plantar sensitivity remained largely unaffected. Together with the behavioural findings, this supports activation of CGRP‑sensitive trigeminal pathways and a migraine‑like phenotype characterised by both light aversion and cephalic hypersensitivity. Importantly, these effects were absent or inconsistent when animals had undergone prior testing or mixed housing, again highlighting the influence of stress and experimental history on outcome measures.

Complementary anxiety testing further strengthened the interpretation of our model. While the open field test proved highly stressful for the animals and difficult to interpret, the dim/dark box provided a calmer and more informative control condition. The absence of anxiety-like behaviour in a low‑light but open environment supported the conclusion that the observed behaviour in the light/dark box reflected light‑specific aversion rather than general anxiety.

Beyond the Results: What This Work Teaches Us About Experimental Design

Taken together, our results demonstrate that intravenously administered CGRP induces light‑aversive behaviour and cephalic allodynia in both male and female rats when tested in an optimised and carefully controlled experimental framework. This is the first preclinical rodent study to combine i.v. CGRP provocation with behavioural and sensory readouts in a way that closely parallels clinical protocols. The work highlights that refinement of classic behavioural models can substantially improve both translational relevance and animal welfare, while also revealing how easily subtle design choices can influence experimental outcomes.

Looking ahead, the combined use of the optimised light/dark box and the dim/dark box provide a promising platform for exploring light aversion and evaluating compounds targeting photophobia. More importantly, the process of developing this model highlights a broader point: robust and interpretable behavioural outcomes depend not only on the intervention itself, but on the many experimental choices surrounding it. Throughout this work, we encountered and addressed challenges related to housing, test design, stress, and prior exposure—factors that initially obscured the effects we aimed to study. These experiences highlight that careful refinement and critical evaluation of experimental design, sometimes through simple adjustments, are essential for producing reliable and translationally meaningful results.

From a personal perspective, as veterinarians, this project was particularly meaningful. Working to refine a long‑established behavioural assay allowed us to align scientific robustness with improved animal welfare, while also gaining deeper insight into how sex differences and social context shape behavioural readouts. These lessons will continue to inform how we approach experimental design in future studies.