We can't read the engine's mind. The evaluation is the result of calculation to depth 22, i.e. looking 11 moves ahead; the main line(s) after both 1 Bxe4 and White's alternatives should give some idea of why the engine gives White +0.6 here and why the alternatives (mainly Bg4 and Bh3) are worse.
Looking at it without an engine, we might guess that after the trades on e4 White has a worse Bishop but more space and a Queenside majority -- and no more worries about a Black attack on the Kingside. After 1 Bxe4 Qxe4 2 Qxe4 dxe4 White can move the Bishop outside the pawn chain with tempo (discovered attack on bPe4), and 3 Bg5 also controls d8 so that after Black defends the pawn White can continue 4 Rd1 and 5 d5, curing the d5-hole (White doesn't want to see Black play Rd5 or Kd5), start getting the Queenside pawns moving, and maybe use the d-file. White also has the option of not trading Queens and instead trying to use the e-file with something like Qd1 (Black won't play Re8 because then Bd2 wins).
If Black answers 1 Bxe4 with dxe4 then White seems to win the e-pawn by moving the Bishop (2 . . . f5 doesn't help Black because 3 f3 and the pinned pawn falls); then the "+0.6" might mean that Black has some compensation thanks to the better Bishop.