keep in mind reductions with sodium are usually done as a mercury sodium amalgam, usually 5% sodium in mercury w/w. And I believe that is what they are referring to in this procedure.
Reducing the fully aromatic beta carboline is not easy, not even LAH would do it. And it's not a practical reagent either and shouldn't even be discussed.
sodium metal reductions work because it occurs through a different mechanism, its a single electron transfer, not a hydride source. It's similar to a birch reduction.
Hydride sources can work if the imine nitrogen of the beta-carboline is alkylated to form the iminium ion, but then you wouldn't have THH when reduced.
Harmaline can be easily reduced with borohydride as I demonstrated in a thread
here on the forum, I also tested neat sodium in ethanol, without the amalgam. The reaction wasn't even partially complete, the amalgamated sodium is too reactive and just reacts with the ethanol before it can react with the harmine/harmaline. Perhaps it's possible though, keeping the solution very cold and with a very slow addition of sodium. I only did a small test.
Fischer might have done it this way in 1889, I don't know for sure if he used amalgam or not. I can't find a copy of the journal (Chem. Ber. 1889, 22, 637)
tert-amyl alcohol is better its slower to react with sodium, he uses in his second prep. Perhaps sodium and isopropanol is worth trying.
Either way, sodium reductions are not really kitchen safe, addition of sodium to ethanol generates enough heat to ignite and explode if not carefully controlled.