It is important to remember that genetic data can conflict or mislead. In particular the matrilineal (mtDNA) and patrilineal (Y-chromosome) signatures may point in very different directions. Even if the per-generation difference in migration or mate selection between males and females is small, the effect will be amplified exponentially over the 300 generations separating us from the first farmers.
I was disappointed to see discussion dwindle, as the topic is fascinating. If the fault is mine for not being controversial enough, let me try to make amends! 
The only pre-IE language in Europe about which much is known is Basque (Etruscan speakers in Italy may have arrived in the Early Iron Age, long after the arrival of Italic speakers.) The coastal regions of Spain, S.W. France, and especially Portugal provide excellent habitat for hunter-gatherers; genetic evidence suggests that it is from here that the Magdalenian culture spread toward North Central Europe as glaciers retreated. This is shown by some genetic markers which are especially strong among the Basque, centered where the Magdalenians presumably originated.
BUT the Y-chromosome picture is quite different. The mutation that dominates Western Europe today didn’t even exist until thousands of years after the Magdalenian expansion. The male and female lines of Western Europe had completely different sources. My previous quote from Cunliffe hints at this:
The almost-complete replacement of Western Europe’s Y-chromosome makes one suspect there was brutality involved, with male invaders enjoying their conquests.
As for language – I think it is normal for a people to switch to the prestige language of their male conquerors; attempts to relate Basque (or, worse, Indo-European) to pre-Neolithich European languages are doubtful. And there is another reason to suspect the proto-Basque language was brought to Spain by its Neolithic conquerors:
Linguists like John Bengtson and George Starostin claim a genetic relationship between Basque and the Northern Caucasian languages; based on cognate counts, Starostin dates their mutual ancestor to 7000 BC, just right if they both derive from the early Anatolian farmers. (I realize that such dating is considered extremely problematic, but the basic idea shouldn’t be in doubt – just the size of the “error bars.”)
The claimed connections between Basque and Caucasian languages are very controversial, and probably rejected by a majority of linguists. My own layman’s impression is that the “splitters” tend to get overly hysterical! :dubious: Here is one paper on the topic showing not only cognates, but hints at regular sound changes. (This paper seeks to relate the minor Burushaski language of Pakistan to Vasco-Caucasian, so shows only three-way cognates.) Here’s an example cognate trio from that paper:
[ul][li] zikiro (castrated ram, Basque)[/li][li] ts’ts’ik’er (young goat, Karata of N.E. Cauc.)[/li][li] tshigir (she-goat, Burushaski)[/li][/ul]
This is one of many examples; can the sound resemblance be mere coincidence? Such cognates for domesticated animals (and others for cultivated crops) suggests, as with Starostin’s date estimate, that any connection is Neolithic.