When Bill Langston claimed that his 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated monkeys were the first model of Parkinson’s disease, he did not really mean that he was ignorant of all the work till then. The problem with earlier rodent models of the disease was their lack of “face validity”. The animals did not have the classical clinical symptoms of the disease. The monkeys had profound bradykinesia, which could be unilateral if the toxin was limited to one side of the brain. They had altered gait and sleep and even developed tremor in some cases (1).

In the earlier models, rodents were dead in less than a week if the lesions were bilateral (2) and if not, then they tended to turn one way if disturbed and to run in tight circles in response to drug treatments (3). Those circling behaviours were instrumental – are instrumental – in developing drugs capable of stimulating dopamine receptors that have had palliative effects in patients. However, the rats did not have the classical symptoms; they were a model of the end-stage neuropathology and could be useful in examining that end stage (4, 5), but the behaviour of a four-legged rodent was just not similar enough to two-legged primates, for the analogy to be obvious.

Even in the days before MPTP, there was an awareness of that problem. Turning rats were good test systems for drugs. They were excellent as models of replacement strategies because they could be transplanted with replacement cells. The behavioural “recovery” was easy to measure – the drugs resulted in the rats turning in the opposite direction to that before the transplant ((6) and many more before that). But did the rats have any bradykinesia?

One of the tests of motor dysfunction did indeed show one kind of bradykinesia. The animal is held in the experimenter’s hand, including both hind legs. By watching carefully how the front legs are used in stepping (or in the “placement reaction” as some surface presents itself within the reach of the forepaws), it is easy to see, even without drug treatment, that the animals are abnormally bradykinetic on the side contralateral to the damage of the dopamine neurons (7).

When I asked Daniel Ogura-Okorie to train rats to use one forepaw in an operant task, he said, “I don’t need to train them, they always use the same paw!” We found that most of the animals were “left handed”. What was obvious was that as soon as hungry animals learned to press a lever for food, they quickly preferred to do it with one forepaw. Following a 6-hydroxydopamine (6-OHDA) lesion of the brain on the side opposite to the preferred paw, they stopped responding, but slowly recovered their learned behaviour – now with the unpreferred paw (8). My first reaction was “Good – now we have a lesion that produces bradykinesia. Our rats are more like patients than we thought”. We could show that the animals were only incapacitated if the lesion disrupted the use of their preferred paw (it did not matter which was the preferred) but that the behaviour reverted back – the animals used the previously unpreferred paw to press. The recovery was so complete that I leaned towards the idea that here was an effect of dopamine that was “purely motor”.

Two things were a little puzzling. Firstly, if we made the lesion and then trained them to press, they did exactly what we would have predicted from our analysis – they learned to press normally, but always used the paw not affected by the lesion (i.e. ipsilateral to the lesion). Then, I had the idea of trying to train them to use the “affected bradykinetic” paw (i.e. contralateral to the lesion).

In stark contrast to the usual 10-min learning curve, if I insisted that the rats were only able to get food if they used the “bradykinetic” paw, I could not get some of them to press at all. After several days of training, the “clever” ones would lean against the wall above the lever and slide down it so that the disabled paw hit the lever. I guess I was soft enough to count that as a success, though compared to the way animals usually press levers, in an operant box, this was not really a “press”. All this was extremely time consuming; I needed to watch the animals throughout the training session.

The first problematic result was that the rats did not learn a new behaviour involving the “affected bradykinetic” extremity (i.e. contralateral to the lesion) even though they did use it to walk – and run in circles. I had spent many happy hours lying on the floor in Stockholm making sure that they used all four paws when running in circles under the influence of apomorphine, for instance. Therefore, learning seemed affected after all, even in the “purely motor” context.

The second problem was pointed out to me by Marianela Garcia-Munoz. She was puzzled that the animals took a week to make the switch between paws after the lesion of the dopamine system: the experiment that proved her point she did with Margot Hamilton in the same boxes that Daniel had used. They trained normal rats and just anaesthetised their preferred paw. Now the switch was instantaneous. The animals could not use the paw, they would walk three legged with the forepaw raised and they instantly switched to the other paw to press the lever. Seven weeks later – at a time when the 6-OHDA-treated animals are still using their previously non-preferred paw – these rats returned to their preferred paw, the effect of the anaesthetic having been long gone (9).