This paper in Nature is fascinating - unfortunately, the chemical studies described were not performed on ancient East Asians, but it lines up with archaeological and anthropological evidence worldwide.
There have only been two studies of Palaeolithic modern humans, Homo
sapiens sapiens. A study of the isotope values of humans from the late
Upper Palaeolithic (ca 13 000 years old) site of Gough's and Sun Hole
Cave in Southern England (Richards et al, 2000a) indicated, again by
the delta15N values, that the main source of dietary protein was
animal-based, and most likely herbivore flesh. The second study
(Richards et al, 2001) was a survey of isotope values of humans from
Gravettian and later (approximately 30 000-20 000 years old) Eurasian
sites. The delta13C and delta15N values here indicated high animal
protein diets, but the type of animal protein was more varied than the
Neanderthals, incorporating aquatic foods in their diets. As this
study was a survey, and associated faunal delta13C and delta15N values
were not measured, it is not possible to further pinpoint the sources
of dietary protein at all of these sites. Interestingly, this
adaptation to aquatic resources becomes more extreme in much later (ca
10 000-5000 BP, depending on area) Mesolithic periods in parts of
Europe. For example, isotope studies of Mesolithic humans from the
Danube Gorges in Southeastern Europe indicate that the majority of
protein was from freshwater fish, which is supported by the
archaeological evidence of fishing equipment and large numbers of fish
bones (Bonsall et al, 1997).
More recent archaeological chemical analyses, such as the one done on remains from Tianyuan cave, also find extensive freshwater fish consumption, indicating things were similar in East Asia. The evidence is that the diet of modern humans in the paleolithic worldwide was primarily animal flesh, supplemented by easily gathered plant material.
Paleolithic tools used in the gathering or preparation of plant foods are either absent, or unrecognizable as such - in light of such an absence, and with the evidence that the diet was primarily meat-based, it must be inferred that plant-based foods that required processing or extensive effort to gather was not a large part of the diet.
This includes wild rice and other grains, most of which required extensive domestication efforts. The earliest evidence of rice consumption only dates to the early neolithic, 11-12kybp, and wild barley only goes as far back as 23kybp, but not as a staple, and not in East Asia. There is one 2009 study that concludes that other wild grains were harvested as early as 90kybp, and claims to have found stone tools to prove it, but this is not yet corroborated, and may not be widespread. The chemical analyses show that animal protein was the dominant dietary staple.