news and views

nature neuroscience • volume 1 no 4 • august 1998 265

within and around the activated area hasto be rearranged to restrict the increasein flow to the area of increased activityalone. Penetrating arterioles are criticalfor flow distribution to the cortical lam-inae. Pericytes are abundant at themicrovascular level and are likely to besignificant in the redistribution of localcapillary flow. Because dopaminergicinnervation is in close contact with bothpenetrating arterioles and pericytes,these terminals are well suited to regu-late microvascular flow. Therefore, oneof the functions of the dopaminergicinnervation might be to focus the vas-cular response to the activated regions.

At the same time, intrinsic dopamin-ergic terminals contacting cerebralendothelial cells could modulate thetransfer of substrates across the blood-brain barrier. The cerebral endotheliumis the site of the blood-brain barrier andis important for controlling the home-ostasis of the neuronal microenviron-ment. For example, during neuralactivity, increases in blood-brain-barri-er permeability might facilitate thetransfer of nutrients into the brain andthe removal of metabolic waste generat-ed by active cells. Raichle and colleaguesfirst provided evidence that centralnoradrenergic terminals modulate thepermeability of the blood-brain barri-er9. Thus, dopaminergic innervation,like other catecholaminergic innerva-tion, could modulate blood-brain bar-rier permeability during neural activity,thereby enhancing substrate delivery tothe working brain. This possibilityseems plausible, considering thatdopamine activates adenyl cyclase incerebral endothelial cells, a signal-trans-duction system involved in endothelialcell transport10.

The origin of this dopaminergic vas-cular innervation has not yet been identi-fied. However, the most likely source aredopaminergic neurons in the substantianigra and ventral tegmental area in themesencephalon. The axons of these neu-rons constitute the mesocortical dopamin-ergic pathway and innervate the cerebralcortex (mainly prefrontal, cingulate andentorhinal cortex) in a highly topograph-ic fashion11. Although the function of thisprojection is not well understood, it islikely to be involved in cognition.Dopaminergic innervation could there-fore contribute to the changes in cerebralblood flow associated with cognitive tasksinvolving the prefrontal cortex. This spec-ulation is supported by studies demon-strating alterations in neocortical

2. Iadecola, C. in Cerebrovascular Diseases (edsGinsberg, M. D. & Bogousslavsky, J.) 319–332(Blackwell Science, Cambridge,Massachusetts, 1998).

3. Iadecola, C. Trends Neurosci. 16, 206–214(1993).

4. Edvinsson, L., MacKenzie, E. T. & McCulloch,J. Cerebral Blood Flow and Metabolism (RavenPress, New York, 1993).

5. Cohen, Z., Bonvento, G., Lacombe, P. &Hamel, E. Prog. Neurobiol. 50, 335–362(1996).

6. Vaucher, E. & Hamel, E. J. Neurosci. 15,7427–7441 (1995).

7. Sun, M. H., Ishine, T. & Lee, T. J. Eur. J.Pharmacol. 334, 165–171 (1997).

8. Sharkey, J. & McCulloch, J. in NeuralRegulation of Brain Circulation (eds Owman,C. & Hardebo, J. E.) 111–127 (Elsevier, NewYork, 1986).

9. Raichle, M. E., Hartman, B. K., Eichling, J. O.& Sharpe, L. G. Proc. Natl. Acad. Sci. USA 72,3726–3730 (1975).

10. Bacic, F., Uematsu, S., McCarron, R. M. &Spatz, M. J. Neurochem. 57, 1774–1780(1991).

11. Goldman, R. P. Adv. Pharmacol. 42, 707–711(1998).

12. Taylor, S. F. Schiz. Res. 19, 129–140 (1996).

13. Drevets, W. C. Annu. Rev. Med. 49, 341–361(1998).

14. Brooks, D. J. Eur. Neurol. 2, 26–32 (1997).

15. Sabri, O. et al. Lancet 349, 1735–1739 (1997).

activation in diseases in which there is adeficit of dopamine12, 13.

The close association betweendopaminergic terminals and cerebralblood vessels raises the possibility that dis-turbances in central dopaminergic neu-rotransmission could alter cerebrovascularregulation and blood-brain-barrier per-meability. For example, cerebral bloodflow and its regulation are abnormal inschizophrenia, depression and Parkin-sonism, diseases in which there is dys-function of dopaminergic pathways12–14.In particular, in schizophrenia, alterationsin blood flow in the frontal lobe often par-allel the clinical manifestations of the dis-ease15. Considering the rich perivasculardopaminergic terminals in the frontal cor-tex, it is likely that alterations in thedopaminergic vascular innervation con-tribute to the vascular disturbancesobserved in these conditions. In addition,it is tempting to speculate that alterationsin neuronal microenvironment resultingfrom the dysfunction in vascular regula-tion and blood-brain-barrier permeabili-ty may worsen the disease process andcontribute to its evolution.1. Raichle, M. E. Proc. Natl. Acad. Sci. USA 95,

765–772 (1998).

Unexpected rewards

Animals (including humans) predict the outcome oftheir behavior, and change their behavior based onthe possible rewards they receive.When the rewardsare different than predicted, there are long-termchanges in behavior, but when rewarded exactly aspredicted, we do not change our behavior. On page304, Jeffrey Hollerman and Wolfram Schultz exam-ine the role of dopamine neurons in reward-basedlearning. Monkeys were simultaneously presentedwith two pictures (shown here), but rewarded withliquid only when they touched a lever below one ofthe pictures (marked `rewarded’). Dopamine neu-rons in the substantia nigra and ventral tegmentalarea were activated during early trials when the mon-keys had not yet learnt to associate the correct picturewith the reward. However, as the monkeys predictedthe rewards more accurately with successive trials,activation was reduced. Moreover, the dopamineneurons were activated when rewarded at unpre-dicted times, and depressed when not rewarded aspredicted. The authors suggest that becausedopamine neurons are activated more strongly byunpredicted rather than predicted rewards, they mayplay a role in reward-dependent learning.

Kalyani Narasimhan

Rewarded Unrewarded

© 1998 Nature America Inc. • http://neurosci.nature.com

© 1

998

Nat

ure

Am

eric

a In

c. •

htt

p:/

/neu

rosc

i.nat

ure

.co

m