718 NATURE | VOL 424 |14 AUGUST 2003 | www.nature.com/nature

news feature

Do you want to work here?The architects behind a new generation of laboratories believe theirdesigns can stimulate good science. Laura Bonetta finds out how,and looks at research that may one day help to test their claims.

Perhaps more importantly, at least in thelong term, architects are also thinking abouthow they can test the claims that they makefor the buildings. What does it mean to saythat a building enhances creativity? Can thisassertion be evaluated? A collaborationbetween architects and neuroscientists,established this spring, is aiming to answersuch questions. Work is at an early stage, butthose involved hope eventually to under-stand how architecture — both good andbad — affects our bodies and minds.

The current attention to good lab archi-tecture has it roots in the Salk Institute.Afterhis success with the polio vaccine, Salk decided to create an independent centre for biological research, and believed thatarchitecture could play an important part inmaking it a success. During the late 1950sand 1960s, he worked with Louis Kahn, oneof the twentieth century’s most influentialarchitects,to design an environment that wasboth welcoming and inspiring. “The Salkwas the starting point for using the buildingto support the scientific community ratherthan just provide space,”says McGhee.

Before then, labs were treated in much thesame way as classroom buildings, with a cor-ridor and often windowless rooms down

In 1992, the late virologist Jonas Salk wasasked to accept an award from the Ameri-can Institute of Architects for the institute

he founded in La Jolla, California. In hisspeech, Salk told how he had visited an abbeyin Assisi, Italy, while working on a polio vac-cine early in his career. Although he did notgo into details, he claimed that somethingabout the architectural experience helpedhim to overcome difficulties in his work. In1954, his vaccine underwent successfulnational trials in the United States, and even-tually helped to cut the number of poliocases in the country by tens of thousands.

Are such stories one-offs, or can the physical environment really influence scien-tific creativity? Laboratory designers areincreasingly banking on the hope that it can.Some are trying to maximize the chanceencounters between scientists that spark newdirections in research. Others are designingbuildings intended to operate like real neigh-bourhoods, where workers experience asense of community within a large building.“A lot more thought is going into the archi-tecture of scientific buildings,” says RobertMcGhee, who has been resident architectwith the Howard Hughes Medical Institute(HHMI) for more than 30 years.

each side. The Salk Institute was a departurebecause Kahn had a high regard for naturallight and coupled it with an understanding ofthe functional requirements of a lab. Theouter four walls of the laboratory floors areglass, creating an open and airy environment.Levels containing equipment such as cen-trifuges and glassware are located in floorsbetween the labs,freeing up space for researchand giving scientists room to adapt theirworking environment to new experiments.

Pride of placeEqually importantly, the Salk Institute isarchitecturally adventurous, and is regardedby many as a stunning building. It is a sourceof pride for those who work there, an effectthat university officials are now taking intoaccount when commissioning buildings thatthey hope will attract the best scientists (seeNaturejobs, page 858). “Everyone who worksat the Salk has an intuition that it is a specialplace,” says Fred Gage, a neuroscientist at theinstitute. “The design is externally aesthetic. Itis very hard to pin down, but it has an effect.”

Many science buildings designed sincehave mimicked these design features.Take theHHMI’s Janelia Farm facility, for example.The new multidisciplinary research institute

FRA

NK

GE

HR

Y

© 2003 Nature Publishing Group

is being built into a hill on the banks of thePotomac River near Leesburg,Virginia,and isset to open in 2005. Labs at Janelia Farm arepositioned along one side of a glass corridorthat runs the length of the undulating build-ing.The glazed walls and ceiling will establisha visual connection to the landscape, whilealso providing access to the open terraces andoffices that alternate along the length of theplan opposite the labs. “It is totally differentfrom being in a box,” says McGhee, who hasworked on the design with New York-basedfirm Rafael Viñoly Architects.

In recent years, architects have begun topay attention to a newer concept — the needfor interactive spaces in which scientists canmeet and talk to each other. The trend hasbeen fuelled partly by the rise of fields such as nanotechnology, which require scientists from different disciplines to work together.Researchers who have traditionally been scat-tered across a university campus are nowbeing brought together under a single roof.“People who would not normally speak areput in an environment that stimulates inter-actions,” says James Broach, a former actingdirector of the Lewis-Sigler Institute for Inte-grative Genomics at Princeton University,New Jersey.“The hope is that sparks will fly.”

Most architects agree that vertical dis-tance is more of a barrier than horizontal dis-tance when it comes to researcher interac-tion. Atria and open staircases, which createsight lines between floors, are one solution.In the Max Planck Institute of Molecular CellBiology and Genetics in Dresden, Germany,built in 2001, the atrium occupies the fullheight and width of the four-story building.Heikkinen-Komonen Architects, based inHelsinki, Finland, designed the building sothat suites of labs and offices radiate from theatrium and are connected to its central stair-case by concrete bridges, which the scientistscall Ponte Vecchios, after the famous bridge inthe Italian city of Florence.A café and diningarea, overlooking a rear garden, occupy theground floor. “The mission was to have asmany chance encounters as possible,” saysKai Simons, the institute’s executive director.

A similar approach has been taken byRafael Viñoly’s firm in its work on the Cali-fornia NanoSystems Institute at the Univer-sity of California, Los Angeles, which is dueto be completed at the end of next year. Thearchitects faced a particular challenge at thissite, as the three levels of laboratory spacewill be built around an existing car park. Toensure that the parking areas do not impair

news feature

NATURE | VOL 424 | 14 AUGUST 2003 | www.nature.com/nature 719

movement between the labs, the designincorporates a series of zigzagging rampsthat connect across and between floors.

But creating a building that increasesinteraction between researchers is notenough on its own. Scientists also need spaceand tools to help them bounce ideas off eachother. With this in mind, architects designspecial areas, equipped with whiteboards,that should give researchers comfortable dis-cussion space away from the bench.Decidingwhere to put them, however, can be tricky.“They work very well if they are in the rightplaces,”says McGhee.“But if you put them atthe end of buildings, for example, they areused to get away rather than interact.”

Traffic managementAt the Stowers Institute for MedicalResearch in Kansas City, designed by Peck-ham Guyton Albers & Viets, also based inKansas City, and completed in 2000, thecommon rooms are in areas of high traffic— for example, where shared equipmentsuch as imaging hardware is housed. “Labsoften have the equipment across the hall,”says the institute’s scientific director, RobertKrumlauf. “At the Stowers you have to go toshared areas.” In a similar vein, the Genen-tech Hall building, which opened in Janu-ary 2003 at the new Mission Bay campus ofthe University of California, San Francisco(UCSF), uses staircase landings to fosterdiscussions. “If you are waiting at the eleva-tor or see someone as you are using thestairs, you can flag them down and sit tohave a conversation,” says Keith Yamamoto,whose genetics group moved into the build-ing when it opened.

Food and drink also help to promote dis-cussion, as long as the environment is right.In 1994, the Imperial Cancer Research Fundin London,now part of Cancer Research UK,added a pub to its canteen area — but staffchose to continue using other local venues,because the pub offered no added conve-nience or incentive for socializing, and it wasremoved.Other similar initiatives may proveto be more successful. At the Lewis-SiglerInstitute, which opened last autumn, a criti-cal focal point is the coffee shop, which haswireless Internet connection, in the build-ing’s atrium. “People gather there, “ saysBroach.“It is an opportunity to see someoneyou would not normally see.”

But big, multidisciplinary buildings alsocome with problems. In a centre thatemploys hundreds of scientists, it is easy tofeel lost. To avoid this, some architects aremodelling lab layouts on real cities, with‘street’ areas for interaction and ‘neighbour-hoods’ where researchers can retreat intosmaller groups. “The trend is to think aboutthe design of a building in the way you thinkof cities,” says Steven Wiesenthal, campusarchitect for UCSF.

The scientists involved in designing

Breaking ground: the Salk Institute (top), theCalifornia NanoSystems Institute (top right),the Max Planck Institute’s Dresden centre(right) and the Ray and Maria Stata Center(left), with its informal ‘streets’ (above), are partof an innovative new crop of research buildings.

TO

P,SA

LK I

NST

ITU

TE

;AB

OV

E,F

RA

NK

GE

HR

Y

LEFT

,RA

FAE

L V

IÑO

LY A

RC

HIT

EC

TS;

BE

LOW

LE

FT,H

EIK

KIN

EN

-KO

MO

NE

N

© 2003 Nature Publishing Group

Genentech Hall, for example, borrowed theneighbourhood concept that had been per-fected in labs at the UCSF’s original Parnassuscampus. “The research towers there were 16stories high,” says Yamamoto of the old cam-pus.“So we renovated each floor and startedthe concept for neighbourhoods.” Researchteams were brought together in an area con-taining labs, offices, meeting rooms, breakareas and shared equipment. The plan was tomake people feel at home in a large building.

Genentech Hall, which was designed byUS architects SmithGroup, with the help ofZimmer Gunsul Frasca Partnership,based inPortland, Oregon, is home to more than 70different labs. Each ‘neighbourhood’, hous-ing four or five research groups, has experi-ment space on the outside and offices in themiddle. Research teams within a neighbour-hood may share common research interests— as well as equipment, meeting rooms andeating areas — but are not necessarily fromthe same departments.

The Ray and Maria Stata Center at theMassachusetts Institute of Technology (MIT)in Cambridge, set to open in winter next year,will also subscribe to the neighbourhood con-cept. The centre’s researchers, spanning adiverse range of fields, from linguistics to arti-ficial intelligence,will occupy a series of areas,each with a two-story lounge at its centre.Thebuilding’s architect, Frank Gehry, devised thescheme to provide a visual and physical con-nection between two levels. In addition, eachneighbourhood has a central elevator core, acentral staircase and some double-height labs.

Proof of principleBut will such innovative designs really bearfruit? Is the extra cost of connecting walk-ways justified in terms of real scientific inte-gration? And how do we know whethernatural light and open spaces get the bestout of researchers? For all the good inten-tions of architects and the scientists whoadvise them, there is no proof that any ofthese tactics actually helps to produce betterscience. “Generally speaking, architects havegood intuitive beliefs and some anecdotalevidence, but little beyond that,” says KevinKampschroer, director of research for theUS General Services Administration (GSA),an organization based in Washington DCthat provides and maintains workspace forfederal employees.

The problem that is now beginning to beconsidered by architects. Norman Koonce,executive vice-president of the Washington-based American Institute of Architects(AIA), discussed the issue in an editorial inthe spring 2003 issue of the AIA’s Journal ofArchitecture, asking: “What would it meanfor architects to move beyond an intuitiveand anecdotal rationale in their design?How much better could we serve our clientsand the public if we understood how theirbrains enable perception of their physical

environment and generate physiologicalresponses to it?”

The San-Diego-based Academy of Neuro-science for Architecture may be the placewhere such questions can be answered. Theacademy was unveiled in May, and aims tosupport research that bridges the gap betweenneuroscience and architecture. One projectthat will soon be under way will use functionalmagnetic resonance imaging to identify thebrain regions that are activated when peoplelook at historic houses, schools and churches,as opposed to contemporary ones. Otherresearch being considered by the academyincludes how lighting and acoustics can affectcognitive activity, and how the brain enablespeople to find their way around in complexbuildings. Although such issues are not specific to lab design, they should produceresults about issues such as lighting that couldpotentially make labs better places to work.

Little is currently known about theseissues, or even about how to study them.“Wehave no clue how design affects the nervoussystem,”says Gage, who has been involved inestablishing the academy’s research direc-tion, and will advise its researchers whenthey begin work. “We don’t even know whatare the right questions to ask.” John Eber-hard,AIA director for research planning anda key figure behind the academy,believes thatit will be 10–20 years before neuroscientistsknow enough about how the brain functionsto begin to address these questions. “We arejust getting started,”he says.

In the meantime, projects with less lofty

goals are under way. The GSA has entered acollaboration with the AIA and scientists atthe US National Institutes of Health to look atthe effects of stress on productivity in theoffice environment. Within the next month,the GSA will recruit about 200 study partici-pants currently working in old-fashionedoffice environments, with features such as lit-tle natural light. The workers will spend timein three different office spaces over a period oftwo years,during which time their stress levelswill be tracked by monitoring indicators suchas heart rate and hormone concentrations.

The study will begin in the workers’exist-ing offices,after which they will move to tem-porary accommodation, where they willcontinue to be monitored. Meanwhile, theiroffices will be renovated so that factors suchas space, lighting, acoustics and temperatureare altered. The final measurements will betaken back at the renovated offices. Duringthis period, the workers will also be askedabout how they feel about aspects of theirnew environments, such as lighting.

“What we are doing is unique,” saysKampschroer. “There have been some stud-ies that have looked at narrow aspects of onequestion or another, but nothing so compre-hensive.”He hopes that this study,and othersplanned by the GSA, will provide designguidelines based on empirical evidence.“Wewill be able to design for reducing stress,”Kampschroer predicts.

Although researchers warn that they maynot isolate a single variable that reduces stressfor all workers, such projects could form thebasis for a proper understanding of how work-ers interact with their environments. Furtherdown the line, such studies could also con-tribute to a thorough evaluation of the way inwhich laboratories are designed,although thisis still many years away. Eberhard draws com-parison between architects today and medicaldoctors in the 1860s: “They did their best forpatients but did it before anyone knew aboutmechanisms of disease. We think the sameshift in knowledge that happened to medicinecould happen to architecture.” ■

Laura Bonetta is a freelance writer based in Bethesda,

Maryland.

Academy of Neuroscience for Architecture➧ www.neuroscienceforarchitecture.org

news feature

720 NATURE | VOL 424 | 14 AUGUST 2003 | www.nature.com/nature

New trends: glass walls (top) and open commonareas (above) are recent themes in lab design.

RA

FAE

L V

IÑO

LY A

RC

HIT

EC

TS/

HH

MI

DAV

ID W

AK

ELY

/SO

WE

RS

INST

.

© 2003 Nature Publishing Group