Stable neuron numbers from cradle to graveRichard S. Nowakowski*Department of Neuroscience and Cell Biology, University of Medicine and Dentistry of New Jersey–Robert Wood JohnsonMedical School, Piscataway, NJ 08873

The human cerebral cortex is themain organ responsible for ourcognitive abilities, e.g., our abili-ties to use language and math, to

reason and theorize, and to read andwrite articles such as this one, essentiallydefining who we are and what we do, and,most importantly, storing a lifetime ofexperience. The critical cell type underly-ing these functions is the neuron. The hu-man neocortex contains �1010 neurons(1), and, understandably, the question ofhow and when these crucial cells comeinto being is fundamental to both normaland pathological conditions and behaviors.The modern era of understanding thetime of origin of the neurons of this re-markable structure began with the use ofa by-product of the atomic age, i.e., theradioactive nucleotide tritiated thymidine,to sequence the production of the corticallayers in mice (2) by exploiting its abilityto incorporate into the DNA of proliferat-ing cells. In this issue of PNAS, Bhardwajet al. (3) exploit the ability of another by-product of this era, i.e., the dramatic ele-vation in atmospheric 14C levels from theearly aboveground nuclear tests, to incor-porate into DNA (4) to resolve one of thehottest controversies in modern neuro-science. The question is: Are the neuronsof the adult human cortex produced pre-natally and then stably retained for ourentire lifespan? Or are some new neuronsproduced in the adult and then integratedinto the preexisting circuitry of the cor-tex? In essence, the issue is ‘‘How old arethe neurons in our cortex?’’ Are they con-stantly renewed like the cells of the skinor the gut, or must the cortex make do fora lifetime with an initial set of neurons?

This issue was first addressed by Rakic(5), who examined the brains of adult rhe-sus monkeys that had been exposed onceor even multiple times to tritiated thymi-dine from 3 days to 6 years before analy-sis. In neocortex, he found labeled glialcells but no labeled neurons and con-cluded that ‘‘the full complement of neu-rons in the primate central nervous systemseems to be attained during a restricteddevelopmental period.’’ The introductionof BrdU immunohistochemistry as a non-radioactive marker for DNA synthesis inthe brain (6, 7) led to an explosion ofstudies on adult neurogenesis, eventuallyreaching consensus that two areas of theadult mammalian brain, the subgranularzone of the dentate gyrus and the subven-tricular zone lining the lateral ventricles,continue to produce neurons destined for

the dentate gyrus and olfactory bulb, re-spectively, in adult rodents (8) and to alesser extent in adult nonhuman primates(9). In humans, only the dentate gyruspopulation contributes new neurons (10,11). Continued production of neurons fornumerous other areas of the brain hasbeen suggested, but the correctness ofthese suggestions can be questioned for avariety of reasons, including technical as-pects of the BrdU method (12–14). Themost controversial is the suggestion that alarge number of new neurons are contrib-uted daily to the adult primate neocortex(15, 16). However, efforts to replicatethese results have not been successful ineither primates (17, 18) or rodents (19),and the issue of possible adult neurogen-esis in the neocortex has been vigorouslydebated (12, 20, 21).

A Two-Pronged ApproachBhardwaj et al. (3) address this key issuefor the human cortex using two ap-proaches to determine whether there is (i)continuous production of a long-livedpopulation of new neurons over the lifes-pan of the adult human, or (ii) a produc-tion of new neurons in the adult that haveonly a transient existence.

The first method used by the authors(3) is an ingenious one. They exploit thefact that the aboveground nuclear tests inthe 1950s and early 1960s produced a dra-matic elevation of 14C levels in the atmo-sphere. Since the cessation of theaboveground tests, the levels of 14C in theatmosphere have fallen exponentially be-cause of equilibration with the oceans,dispersion in the biomass, etc. Becausemetabolically active 14C levels in plantsand animals are in equilibrium with atmo-spheric 14C and the half-life of 14C is long(�5,000 years), any molecules synthesizedduring the period of rapidly changing 14Clevels can be dated by the proportion of14C they contain. DNA in nonproliferatingcells is stable, so 14C levels in DNA re-flect the level of atmospheric 14C at thetime of DNA synthesis (i.e., at the time of

cell ‘‘birth’’). This is the essential factthat the authors have grasped and ex-ploited. They reasoned that if the neo-cortical neurons were produced duringthe developmental period, then the ageof the DNA in these neurons would bethe same as the individual (i.e., it woulddate to the approximate birth date ofthe human from which it was taken).They have validated this approach in avariety of tissues (4, 22), and now theyapply the technology in individuals born(in northern Europe) before, during,and after the above-ground test periodusing four different regions of the neo-cortex: prefrontal, premotor (frontal),parietal, and temporal.

The oldest individual analyzed was 72years old and, thus, was �20 years oldwhen the atmospheric levels of 14C startedto rise. Five additional individuals wereborn during or before 1953, the year ofthe first aboveground nuclear tests andthe onset of the rise in atmospheric 14C.The tissue was collected from autopsiesperformed in 2003–2004; thus, the resultsfrom these specimens effectively integrate�50 years of neuron production, i.e., overmost of the expected lifespan of a human.The effect of this integration is that itmakes the analysis quite sensitive to thecontinuous production of even a smallnumber of cells. Both neurons and non-neurons, distinguished by cell sorting, andan unsorted control were analyzed. Theresults show that the average age of theneurons (with respect to the age of theindividual) is age 0.0 � 0.4 years, i.e., thesame as the age of the individual. In con-trast, the nonneuronal cells have an aver-age birth date of 4.9 � 1.1 years after thebirth of the individual. Importantly, theunsorted samples do not differ from thenonneuronal samples, which makes sensebecause glia outnumber neurons by �10:1in the adult. The dating of the neurons tothe same time as the birth of the individ-ual argues strongly that there are no newneurons produced after birth; however,given the precision of the dating methodsused (22), it is estimated that the limit onthe possible number of new neurons madeduring this 50-year period is �1% of thetotal (22). In other words, �99% of the

Conflict of interest statement: No conflicts declared.

See companion article on page 12564.

*E-mail: rsn@umdnj.edu.

© 2006 by The National Academy of Sciences of the USA

In humans, onlythe dentate gyrus

population contributesnew neurons.

www.pnas.org�cgi�doi�10.1073�pnas.0605605103 PNAS � August 15, 2006 � vol. 103 � no. 33 � 12219–12220

CO

MM

EN

TA

RY

human neocortex is produced duringthe fetal period.

To settle the issue of the possible exis-tence of a transient population of newneurons that would not be detected by the14C method, Bhardwaj et al. (3) did a sec-ond experiment using autopsy materialcollected from cancer patients who hadreceived injections of BrdU (10) for diag-nostic purposes from 4.2 months to 4.3years before death. They found BrdU-labeled cells in the neocortex, but none ofthe labeled cells were neurons or neuron-like. As a positive control, they examinedthe dentate gyrus of the same patients,where they found clear evidence of neu-ron production. From this second experi-ment, they conclude that if the neocortexcontains new neurons with a short lifes-pan, those new neurons survive for nomore than �4 months, i.e., the length ofthe shortest survival span they examined.Importantly, the comparison with the den-tate gyrus indicates that any transient newneurons in the neocortex must be consid-erably less prevalent than they are in thedentate gyrus, where �22% of the BrdU-labeled cells apparently have a neuronalphenotype (10). In addition, their resultsmean that transient cells do not persist inan undifferentiated state and later be-come neurons, at least within the �4-yearspan examined.

Functional ConsiderationsBoth of the experiments of Bhardwaj et al.(3) indicate that there are no new neu-rons, either long-lived or transient,produced in the adult human for the neo-cortex. Importantly, these experiments arequantitative and indicate a theoreticalmaximum limit of 1% on the proportionof new neurons made over a 50-year pe-riod. This proportion provides a maximumlimit as to the possible rate of production,i.e., 0.02% per year. With respect to thefunction of the neocortex, this maximum

rate of production should be consideredin the context of the cortical column, i.e.,the physiologically functional unit of infor-mation processing in the cortex (23, 24).In the neocortex, the minicolumn, alsoknown as an ontogenetic column (25), isthe fundamental anatomical and process-ing unit that is iteratively arrayed acrossthe neocortical surface; each minicolumncontains 80–100 neurons but may be 2.5times that size in the visual cortex (23,24). The crucial relationship here is that,

regardless of the size of the cortical col-umn, the estimated limit on the numberof new neurons produced is extremelysmall. At the hypothetical maximum limitsof production, the larger columns in thevisual cortex might receive one new neu-ron every 20–30 years, whereas thesmaller columns in the rest of the neocor-tex might receive one new neuron every50 years. In other words, even the maxi-mum theoretical production of new neu-rons is so small as to make clear that thepotential contribution of new neurons toneocortical function is minuscule to non-existent. In contrast, the BrdU experimentof Bhardwaj et al. (3) indicates that thereis a constant production of new glial cellsin adults. This finding is consistent withtheir 14C data that reflect an average birthdate of �5 years after the birth of theindividual for the nonneuronal cells. Thefact that there is continuous glial cell turn-over might be important for pathologyand treatment of injury (13).

Stability ReignsBhardwaj et al. (3) settle a hotly contestedissue, unequivocally. The two-prongedexperimental approach clearly establishes(i) that there is little or no continuousproduction of new neurons for long-termaddition to the human neocortex and (ii)that there are few if any new neurons pro-duced and existing transiently in the adulthuman neocortex. Importantly, the resultsare quantitatively presented, and a maxi-mum limit to the amount of production ofthe new neurons can be established fromthe data presented. The data show thatvirtually all neurons (i.e., �99%) of theadult human neocortex are generated be-fore the time of birth of the individual,exactly as suggested by Rakic (5), and theinescapable conclusion is that our neocor-tical neurons, the cell type that mediatesmuch of our cognition, are producedprenatally and retained for our entirelifespan. New neuron production in thedentate gyrus may have a hypotheticalrole in formation of certain types of mem-ory (26), but its absence in the neocortexpenetrates to the essence of how we thinkand learn. If neuron number in the neo-cortex is not incremented, then synapticchanges and other forms of plasticity mustdominate; i.e., when we learn a new taskor fact, then the storage of the new infor-mation must entail a reorganization ofexisting circuitry. The retention of theneuronal population for decades is consis-tent with our need to retain informationfor our long lifespan. In short, the culturalcomplexity of humans requires not onlythe constant acquisition of new facts andskills but also the retention of others,most notably language, for many decades,and a stable complement of neurons inthe neocortex would seem to be essentialfor these abilities. This could be a reasonfor the evolutionary choice for a stablecellular composition of our cognitive ma-chinery over a more dynamic pattern.

1. Blinkov, S. M. & Glezer, I. I. (1968) The HumanBrain in Figures and Tables: A Quantitative Hand-book (Plenum, New York).

2. Angevine, J. B. J. & Sidman, R. L. (1961) Nature192, 766–768.

3. Bhardwaj, R. D., Curtis, M. A., Spalding, K. L.,Buchholz, B. A., Fink, D., Bjork-Eriksson, T.,Nordborg, C., Gage, F. H., Druid, H., Eriksson,P. S. & Frisen, J. (2006) Proc. Natl. Acad. Sci. USA103, 12564–12568.

4. Spalding, K. L., Bhardwaj, R. D., Buchholz, B. A.,Druid, H. & Frisen, J. (2005) Cell 122, 133–143.

5. Rakic, P. (1985) Science 227, 1054–1056.6. Miller, M. W. & Nowakowski, R. S. (1988) Brain

Res. 457, 44–52.7. Nowakowski, R. S., Lewin, S. B. & Miller, M. W.

(1989) J. Neurocytol. 18, 311–318.8. Alvarez-Buylla, A. & Lim, D. A. (2004) Neuron 41,

683–686.

9. Kornack, D. R. & Rakic, P. (1999) Proc. Natl. AcadSci. USA 96, 5768–5773.

10. Eriksson, P. S., Perfilieva, E., Bjork-Eriksson, T.,Alborn, A. M., Nordborg, C., Peterson, D. A. &Gage, F. H. (1998) Nat. Med. 4, 1313–1317.

11. Sanai, N., Tramontin, A. D., Quinones-Hinojosa,A., Barbaro, N. M., Gupta, N., Kunwar, S., Law-ton, M. T., McDermott, M. W., Parsa, A. T.,Manuel-Garcia Verdugo, J., et al. (2004) Nature427, 740–744.

12. Nowakowski, R. S. & Hayes, N. L. (2000) Science288, 771a.

13. Ackman, J. B. & Loturco, J. J. (2006) Exp. Neurol.199, 5–9.

14. Kuan, C. Y., Schloemer, A. J., Lu, A., Burns, K. A.,Weng, W. L., Williams, M. T., Strauss, K. I.,Vorhees, C. V., Flavell, R. A., Davis, R. J., et al.(2004) J. Neurosci. 24, 10763–10772.

15. Gould, E., Reeves, A. J., Graziano, M. S. & Gross,C. G. (1999) Science 286, 548–552.

16. Gould, E., Vail, N., Wagers, M. & Gross, C. G.(2001) Proc. Natl. Acad Sci. USA 98, 10910–10917.

17. Kornack, D. R. & Rakic, P. (2001) Science 294,2127–2130.

18. Koketsu, D., Mikami, A., Miyamoto, Y. & Hisat-sune, T. (2003) J. Neurosci. 23, 937–942.

19. Ehninger, D. & Kempermann, G. (2003) Cereb.Cortex 13, 845–851.

20. Gould, E. & Gross, C. G. (2002) J. Neurosci. 22,619–623.

21. Rakic, P. (2002) Nat. Rev. Neurosci. 3, 65–71.22. Spalding, K. L., Buchholz, B. A., Bergman, L. E.,

Druid, H. & Frisen, J. (2005) Nature 437, 333–334.23. Mountcastle, V. B. (1997) Brain 120, 701–722.24. Buxhoeveden, D. P. & Casanova, M. F. (2002)

Brain 125, 935–951.25. Rakic, P. (1988) Science 241, 170–176.26. Aimone, J. B., Wiles, J. & Gage, F. H. (2006) Nat.

Neurosci. 9, 723–727.

There are no newneurons produced

in the adult humanfor the neocortex.

12220 � www.pnas.org�cgi�doi�10.1073�pnas.0605605103 Nowakowski