Histological data: Hollard and Davison (1971).

Keller et al. (1978) looked at pigeon brains after a self-stimulation study. They sliced the brain tissue to see where the tiny electrodes had sat.

The birds had pecked for hours to zap the same spot. The team wanted to know if the electrodes were truly in the planned outer layer, the ectostriatum.

Every electrode had missed the target. The tips sat in the older paleostriatum, not the ectostriatum.

The mistake explained why the pigeons kept pressing: the wrong spot still felt good. Without the brain check, the old data would have stayed misread.

Lattal et al. (2022) argue that every animal-training claim should first pass a basic lab test. Keller et al. (1978) is the kind of careful anatomy check they call for.

Rogers-Warren et al. (1976) removed electrodes entirely by using fur-safe static shock. Their paper and Keller et al. (1978) move in opposite directions—one avoids implants, the other audits them—yet both chase the same goal: clean stimulation data.

Reed et al. (2012) and Hoyle et al. (2022) also double-check their gear, using ERP and ERG to be sure their signals are real. Keller et al. (1978) adds post-mortem histology to that toolkit.

Before you trust any brain-based intervention—whether you work with pigeons, parrots, or people—prove the hardware is where you think it is. Build a quick histology, imaging, or signal-verification step into your protocol. One small check can save months of noisy data and faulty conclusions.