Yesterday I let myself be carried away by the apparent antiquity of Aterian in North Africa. I was already persuaded that Aterian and the arrival of mitochondrial DNA U6 to North Africa was the same thing. True that I had arrived to such conclussions when I thought Aterian had much more recent dates... enfin.
Something I should have re-read before posting that is: N. Maca-Meyer's paper on the philogeny and possible spread scenarios of U6 in North Africa. True that they conclude that the arrival of U6 to North Africa originated in West Asia... but they also notice that the nucleotide diversity of the haplogroup is higher in Iberia than in Africa. The problem? That U6 is seldom seen elsewhere in Europe, so she thinks that the clade arrived to Spain via North Africa and not vice versa. Maca-Meyer thinks that U6 arrived from West Asia via East Africa, based only in the greatest diversity of one subclade U6a in East Africa.
The estimated age of U6 is of c. 66,000 BP (+/- 25,000 years), what certainly could make it coincident with the arrival of Aterian via East Africa (The Horn). Aterian seems indeed most closely related to African Middle Stone Age, like Indian Middle Paleolithic, with dates c. 60-73,000 BP but some reaching as early as 90,000 BP. So yes, in principle, Aterian and U6 correlate well.
Another possibility could be that U6 arrived to North Africa from Iberia with the Oranian (Iberomaurusian) culture, that is believed to be an offshot of Mediterranean Iberian Gravettian or later Gravettizing cultures. Nevertheless Oranian correlates more clearly with haplogroups H and V. If that was the case, it would not be the only case of a U clade that is as old as Aurignacian but is found only rarely in Europe, due to drift. The case of U8a can serve maybe as counter-example. If U6 was accidentally concentrated in Mediterranean Iberia due to founder effect, it could well have expanded to North Africa from there along H, V and possibly other smaller clades without ever migrating to continental Europe, at least in significative ammounts.
This can only be understood if one knows reasonably well the peculiarities of Mediterranean Iberian UP, always reciever and almost never exporter of culture (with the possible exception of some facies of Solutrean, restricted anyhow to the Iberian peninsula). In Aurignacian, Gravettian, Magdalenian and the epi-Paleolithic, Mediterranean Iberia is always at the recieving end, at least in what regards to Europe.
So personally I would not exclude the model of a European (Med. Iberian) origin of U6. If we are not going to push back the ages of the OOA event too much, then it looks the more plausible explanation. Occam's razor seems to favor it, really.
Still the late Paleolithic of North Africa is intriguing. The older ages of Aterian have been reviewed to as early as c. 90,000 BP (Cremaschi et al, 1998) and the end of this culture is uncertain, with dates wildly varying between 65,000 BP (Cremaschi, who contests older C14 dates) to 25,000 or even 15,000 BP (Thillet).
Then you have the Oranian (Iberomaurusian) with an older datation of c. 22,000 BP (source: J. Escola Pujol, on Uadi Kenta - in Catalan), making it coincident with the complex Gravettian-Solutrean transition in Mediterranean Iberia (that culminates with a Gravettizing Solutrean that could well be called a Solutreanized epi-Gravettian as well). The most recent dates could be of c. 7000 BP.
And then you have the more famous Capsian cuture that is clearly epi-Paleolithic, coexisting with Oranian c.10,000-6000 BP, that is of quite clear East African origin (Sudan, The Horn) and that is with all likehood related to the spread of Y-DNA E3b and Afroasiatic languages (Berber).