I'm scratching my head to understand this behaviour of MultiSeries. If I type

MultiSeries[series](diff(ln(GAMMA(1/x)), x), x)

Maple 2022.2 answers

But as is known (and maple confirms)

diff(ln(GAMMA(x)), x)=Psi(x)

But prompting in maple 

MultiSeries[series](Psi(1/x), x)

It answers

that is off by a x^2 factor with the former answer, while one expects to obtain the same answer as before. Why is this? (I tried assuming positivity, same discrepancy results).