Is my INCAR wrong in NEB ??

[matylove]

system=NEB

SPRING=-5
IMAGES=7
IALGO=48
ISMEAR=0
ENCUT=500
PREC=High
ISTART=0
NPAR=1

NSW=200
ISIF=2
IBRION=1
NFREE=10
POTIM=0.5

[graeme]

That looks reasonable. We generally find that high precision is unnecessary, but maybe not for your system. You might also start with IBRION=3 and POTIM=0.1 in case you have high initial forces on the NEB images. Using POTIM=1 should be more efficient in the harmonic basin near convergence. But also take a look at our new optimizers: http://theory.cm.utexas.edu/vtsttools/optimizers/ ; in our tests, these work somewhat better than the versions built into vasp, and in the case of LFBGS significantly better for the NEB.