1、Behavior Research Methods, Instruments, & Computers2000,32 (2),290-296MATLABand graphicaluserinterfaces:Tools for experimentalmanagementERINM.HARLEYandGEOFFREYR.LOFTUSUniversityofWashington, Seattle, WashingtonMATLABis a convenient platform for the development and management of psychological experi-ments because ofits easy-to-useprogramming language,sophisticatedgraphics features, and statisticsand optimization tools. Through implementation of theBrainard-PelliPsychophysics Toolbox, theMATLABuse
2、r gains close temporal and spatial control over theCRT,while retaining the simplicity ofan interpreted language conducive to rapid program development.MATLABsabilities can be furtherutilized through easily programmable graphical user interfaces (GUIs). Weillustrate how a GUIcanserve as a powerful and intuitive tool for organizing and controllingallaspects of a psychological ex-periment, including design, data collection, data analysis, and theory fitting.We describe here three related topics: fi
3、rst, the useofMATLAB in psychological research; second, the valueofgraphical user interfaces (GUIs) inorganizing and runningexperiments; and third, our implementationofone suchGUI within the MATLAB environment.MATLABANDPSYCHOLOGICAL RESEARCHOver the past few years, MATLAB (a productofTheMath Works Inc.) has gained popularity as a meansofrunning psychological experiments, particularly thoseinvolving perceptionandcognition. MATLAB, whichstands formatrix laboratory,is a powerful platform forhigh-pe
4、rformance mathematical computation and graph-ical representation, whose basic data element is anN-dimensional matrix(Hanselman&Littlefield, 1996).MATLABs primary strength as a means for creating andexecuting psychological experiments lies in its compre-hensive packageofsuperior mathematical and graphicaltools. Within this single application, allofthe followingexperimental tasks can be achieved: creation and manip-ulationofstimuli, design and executionofthe experiment,data collection, statistical
5、 analysis, fitofcomplex theoryto data, and interactive graphical data display.MATLAB runs onMacintosh,Windows, and UNIXoperating systems and includes both a programming lan-guage and a large setofmathematically oriented libraries.?MATLAB serves as a good alternative to popular plat-The writingofthis manuscript was supported by NIMH GrantMH4I637 to G.L. We thank David Brainard for enormous amounts ofhelp inunderstanding the Psychophysics Toolboxand MATLAB in gen-eral. The MATLABcode discussed int
6、hisarticle will be posted on E.H.sWebsite (http:/students.washington.edu/eharley/). Please e-mail eitherofthe authors for details.Correspondenceconcerning this articleshould be addressed to E. M. Harley, Department of Psychology, Uni-versityofWashington, Seattle, WA 98195-1525 (e-mail: eharleyu.washington.edu).forms, such as PSYSCOPE (Cohen, MacWhinney, Flatt,& Provost, 1993; Yee & Vaughan, 1999) and MEL(St. James&Schneider, 1991; Schneider, 1989), becauseit provides greater control over experim
7、ental design, al-lowing for the creationofentirely unique experiments.MATLAB also serves as a good alternative to low-levelprogramming languages, such as C and PASCAL, be-cause MATLABs language is abstracted from the hard-ware details and is generally easier to learn. As a result,a MATLAB user receives the benefitsofa programminglanguage that is both conducive to rapid program devel-opment and able to perform with the power and flexibil-ityofalow-level language (primarily through the additionoft
8、he Psychophysics Toolbox, discussed below).ThePsychophysics ToolboxMATLAB is particularly useful for running sensoryand perceptual experiments, largely owing to an exten-sive setoftools-thePsychophysicsToolbox;writtenby Denis Pelli and DavidBrainard-thatallows extremelyclose temporal, spatial, and color-related control overCRT-based stimulus displays (Brainard, 1997; Pelli, 1997).Specifically, the toolbox provides access to the comput-ers display frame buffer and color lookup table, allowssynchr
9、onizationofstimulus display with the vertical re-trace, supports millisecond timing, and facilitates the col-lectionofobserver responses (Brainard, 1997). The tool-box incorporates these features into a large collectionofstimulus-display routines that allow the user to store pre-formed visual images (e.g., scenes, forms, text, or somecombinationofthese) in the computers memory and todisplay these images at rates that are constrained only bythe monitors refresh rate. For example, on a MacintoshG3
10、 computer, a 400X400 pixel gray-scale image can betransferred from memory to the screen buffer in less than10msec, thereby allowing a sequenceofsuch images (thenumberofwhich is limited only by computer memory) tobe displayed at the fastest possible rate.Copyright 2000 Psychonomic Society, Inc.290MATLAB AND GRAPHICAL USER INTERFACES291The Psychophysics Toolbox has been used by re-searchers to study a varietyoftopics in the fieldofpsy-chology, including face and object recognition, psycho-physical
11、 thresholds,colormatching, visual search,categorization, motion detection, and perceptual learning.Later in this article, we will briefly discuss our useofthetoolbox to display high- and low-pass filtered digit stringsfor verybriefexposure durations.Tools for Data Analysis and Theory FittingMATLAB provides a wide rangeoffunctions for useindataanalysis. For example, the optional StatisticsToolbox includes more than 200 routines covering topicsincluding, but not limited to, descriptive statistics,
12、 prob-ability distributions, linearand nonlinear modeling, multi-variate statistics, and hypothesis testing. MATLAB alsoprovides graphing tools for the creationofpresentation-quality plotsofvarious forms, including line, scatter,stem, pie, and box plots and bar graphs and histograms, aswellas three-dimensional contour, surface, and mesh plots(with optional animation).Experimenters in the fieldofperception and cognitionoften wish to fit mathematical theory to their data. MAT-LABs optional Optimiz
13、ation Toolbox provides the toolsnecessary for fitting complex, nonlinear mathematicalmodels with a numberoffreeparameters but no analyticsolution. For a given data set, the optimization routinesfind the valuesofthe free parameters that minimize somedesired error function (e.g., the root-mean square errorbetween the data and the theory predictions).MATLABs Intuitive Programming LanguageMATLAB incorporates a programming language thatis similar in structure to many common languages, suchas FORTRAN,
14、 BASIC, PASCAL, and C. One elementthat makes MATLABs language unique is that, as wasnoted, it uses a matrix as its basic element. This propertyconfers numerous benefits, one being the following: Inmost programming languages, it becomes time consum-ing and cumbersome to perform the same mathematicaloperation on a groupofnumbers, because the operationmust be repeated (generally via loops) for each number,one number at a time. The greater the dimensionalityofthe data structure on which suchoperatio
15、nsare per-formed, the greater the numberofembedded loops re-quired. However, in MATLABs programming language,data structures are construed as matrices, and almost anyoperation, be it as simple as adding 1to each structure el-ement or as complex as finding all elementsofa structurethat are less than 0, can be performed in a single step. Em-bedded loops for these tasks are thus eliminated in MAT-LAB, saving the programmer time and cutting down onthe length and complexityofprogramming routines. Ase
16、cond benefitofMATLABs matrix-based languagepertains to the use and manipulationofimages. Imagesare stored in MATLAB as matrices, and becauseofthis,they can be quickly and easily manipulated or altered (e.g.filtering, adding noise, reducing/increasing contrast, etc.)through matrix operations.Like other programming platforms, MATLABprovidesautomatic formatting features, such as color codingofcomments, strings, and key words, as well as indentationof loops, thus increasing clarity and organizationo
17、fpro-gramming code. The net effectofsuch features, combinedwith MATLABs simple and intuitive programming syn-tax, is to allow someone with little or no prior program-ming experience to quickly become a proficient MAT-LAB programmer and to write sophisticated routineswithin a matter of weeks.Graphical User InterfacesThere are essentially three ways in which a user cancommunicate with the computer via MATLAB: throughthe Command Window, through the useofscripts andfunctions, and through GUIs.TheCom
18、mand Windowis MATLABs default I/O tech-nique. As the name implies, it is a window into which anystandard MATLAB commands or user-defined com-mands, such as 2+2 or answer=conv2 (mask, filter),can be typed. Although the Command Window is ade-quate for accomplishing simple tasks, it is often useful tocreate a file containing a listofsuch commands. Thesefiles are calledscriptsifthey simply run a listofcom-mands andfunctionsiftheyacceptinput argumentsand/or return output arguments. Both scripts and f
19、unc-tions can be executed either directly from the CommandWindow or from within other scripts or functions. Topro-vide an example, a function that we wrote and use oftenis called ComputeContrast.Itrequires two input val-ues, foreground color (FC) and background color (BC),and returns the contrast value calculated by the formula(FC - BC)/(FC + BC).The third communication device, a GUI, provides anintuitive interface between the user and the program-ming language. A GUI allows the user to bypass M
20、AT-LAB commands altogether and, instead, to execute pro-gramming routines with a simple mouse click or keypress(most common applications, such as Word and Photo-shop, as well as the Macintosh and Windows operatingsystems, are implemented as complex GUIs). No knowl-edgeofMATLABor computerprogramming is necessaryfor a user to successfully navigate a well-designed GUI;indeed, from the users perspective, the language under-lying the GUI is irrelevant. GUIs can range from simplequestion boxes prompti
21、ng the user for a Yes/Noresponse,to more complex interfaces, an exampleofwhich will beprovided below.MATLABprovides the user with intuitivetools for easy construction ofGUIs; see Marchand (1999)for a detailed description of the GUI-creation tools pro-vided by MATLAB.USING GRAPHICALUSERINTERFACESIN PSYCHOLOGICAL EXPERIMENTSTo illustrate the benefits of using GUIs in psychologi-cal experiments, we will describe a specific GUI createdto accompany a visual perception experiment in our lab-oratory.29
22、2HARLEY AND LOFTUSExample: The Filter ExperimentTo provide the reader with a foundation for under-standing and interpreting our exampleGut,we will firstbriefly describe the experiment for which the GUI wascreated. The experiment, which we will call thefilterex-periment, was designed to investigate certain kindsofspatial-frequency filtering on perception and memoryforbrieflypresented digit strings (see, e.g., Olds&Engel,1998; Parish&Sperling, 1991). Stimuli were randomlychosen four-digit strings
23、presented in three spatial-frequency conditions: normal (N), low spatial frequen-cies only (LSF), and high spatial frequencies only (HSF).On a given trial, the following sequenceofevents oc-curred. First,a stimulus in oneofthethree spatial-filteringconditions was presented at oneofsix exposure dura-tions; next, the observer attempted to recall the digits intheir correct order (guessing if necessary); and third, vi-sua?and/or auditory feedback was provided. The pro-grams used in the filter experi
24、ment were all written inMATLAB, utilizing standard MATLAB routines in con-junction with the stimulus presentation routines in theBrainard-Pelli Psychophysics Toolbox.Organizing Multiple Data SetsFourteenobservers participatedin the filter experimentand produced a totalof64 data sets (where a data set is asetofexperimental trials, all involving the same experi-mental parameter values). Three observers collectedpilotdata during the developing stagesofthe experiment.Three different observers collec
25、ted data for the final fourversionsofthe experiment (in which mask/ no mask andlow-pass filter size were varied). Finally,8observers col-lected data to serve as practice fora follow-up experimentand todetermine the appropriate experimental parameters(contrasts, durations, etc.) to be used in that follow-up ex-periment. Although it might seem as if organizing allofthese observers and data sets, with their differing experi-mental parameters, would be time consuming and com-plicated, aGutdesigned e
26、xpressly for this experimentallows for simple control overthe multiple data sets fromthe multiple observers.The Example Graphical User InterfaceThe GUI created to accompany the filter experimentconsistsoftwowindows: a main window, pictured in Fig-ure 1, and a theory window, pictured in Figure 2, whichwe will discuss in turn.The GUIs main window contains four major compo-nents: observer information, setup, and stimulus preview;feedback options and data collection; data analysis, plotoptions, and
27、graphical display; and escape from the mainwindow in the formofeither a link to the theory windowor an exit button. Wewill describe eachofthese compo-nents within the contextoftheir contributions to the de-sign, data collection, data analysis, and theory applicationentailed in the filter experiment.ManipulationofExperimental ParametersFrom Within the Graphic User InterfaceLike the filter experiment that we use in our example,many experimental projects, particularly those involvingsensation and p
28、erception, involve a seriesofinterrelateddata sets. For instance, a project to investigate the effectsofcontrast on information acquisition may involve a se-riesofexperiments, each with varying valuesofdurationand contrast and, usually, involving choicesofother ex-perimental parameters as well. Often, such experimentsact aspilotexperiments, designed to map out the layoftheperceptual land for individual observers before the mainexperiment is designed and the final data collected.As thenumberofsuc
29、h data sets within a projectgrows, it becomes increasingly difficult to change theexperimental parameters in a systematic and/or optimalfashion, because such parameters are often embedded inthe codeofthe relevant programs. Keeping trackofwhichdata sets emerged from which observers also becomestedious as the numberofdata sets per observer increases.AGutis a useful tool for dealing with these problems,since it can do double duty as a mini-database. Throughits use, an experimenter can easily set, a
30、lter, and keeptrack of experimental parameters for multiple observerswithin a single visual display.Thedesign features used forthese purposes that are implemented in our example GUI(see Figure 1) are the following.Observerinformation. The area in the upper left sec-tionofour GUIs main window is devoted to observer in-formation and setupofdata files. Here, allofthe ob-servers initials are listed in a pop-up menu. Once aparticular observer has been selected, information forthat observer is loaded,
31、 and theGutisupdated-thatis,parameter values are displayed, and any existing data aregraphed in theGutsgraphical display. If, as is usuallytrue, the observer under consideration has more than onedata set, a specific data set can be called up with the Se-lected Data Set pop-up menu, and again theGutwill beupdated. In Figure 1, observer E.U. is selected; she hasfive existing data sets,ofwhich Data Set 2 has been se-lected for analysis. The various parameter values for thisdata set are shown in the
32、 edit boxes along the left sideofthe GUI. This allows the experimenterto immediately seewhat parameter values characterized Data Set 2 for ob-server E.U.4Setup.Ifthe experimenter wishes to establish a newdata set for an observer, he or she does so by setting allofthe relevant experimental parameters to their desiredvalues and simply clicking on the push button labeledSetup. This button calls a routine that creates a new datafile for the observer with the current parameter settings.An important c
33、larification must be made here about thedual functionofthe GUIs editable parameter list. Aswas just described, one functionofthe list is for the cre-ationofnew data sets. The second functionofthe list istoindicate clearly to the experimenter what parameter val-ues were used in any existing data set. Forexample, whenMATLABANDGRAPHICALUSER INTERFACES293Blocks Run: 24r-ExitPlot linesoNolinesoLogDurationo-lnfl-p)Analyze From Block:ToBlock:n:48;ses=0.032Feedback:oAuditoryIiIVisualRunr-IBothCopyTo Cli
34、pboard.PlotOptions.ObserverIDitials:IED.IExisting Data Sets: 5Durations: 13to147111sContrast: 0.570Low-pass cutoff(cldigit):1.95Selected Data Set:I.ITrialsperBlock:I36Initial Duration(Ticks):I;:=NumberofDurations:I6Durationspacing(ticks)I WindowColor:Stimulus Color:Filter size (pixels):I6IoMaskstimulus?Figure 1.Mainwindow of examplegraphicuserinterface (GUI)createdto accompany anexperiment(thefilter experiment) designed to investigate the effects of spatial frequency filtering on digit recall. N
35、ote: This is ablackandwhite reproductionofamulticolored GUI.a data set is selected for analysis on the GUI, the param-eter values corresponding to that data set are automati-cally loaded and displayed in the editable boxes.The various piecesofinformation below the parameterlist are updated when parameter values are altered. Forexample, if the editable text corresponding to (i.e.,justtothe rightof)NumberofDurations were to be changedfrom 6 to 8, the static text Durations: 13 to 14Twouldchange to
36、read Durations: 13to 200. Note that althoughstatic text cannot be edited on an active GUI, it can bechangedin real time by theunderlyingprogrammingcode to reflect changes made to relevant editable itemson the GUI.Stimuluspreview. The black box in the bottom leftcornerofthe window shows a previewofwhat the stim-uliwilIlooklike during the experiment (showing, fromtop to bottom, N, LSF, and HSF stimuli). Like the statictext below the parameter list, the preview stimuli also294HARLEY AND LOFTUSchang
37、e to reflect the values set in the parameter list. Thisallows the experimenter to see how manipulating vari-ables, such as stimulus color (gray-scale value rang-ing from 1 to 256) or filter size, will affect the appear-ance of the stimuli, without entailing actual executionofthe experiment.Execution of theExperimentfromWithin theGraphicUserInterfaceIn visual perception experiments, it is common to havefew observers but a great dealofdata per observer. Withthe aidofa GUI, observers can easily col
38、lect their datawithout the presenceofthe experimenter.Data collection.In the experiments conducted in ourlaboratory, observers are instructed to open MATLABand to type the nameofthe experiment in the CommandWindow, which calls up and executes the relevant GUI.Once the GUI is open, the observers select their initialsfrom the Observer Initials pop-up menu, select the ap-propriate data set from the Selected Data Set pop-upmenu, and click on the button entitled Run. While theexperiment isrunning, th
39、e MATLAB code automaticallysaves data to the observers data file on completionofeach blockoftrials, thereby eliminating the necessity forthe observers to be responsible for saving their own datato a file atthe endofa data collection session. At the endofa session, observers are returned to the GUI, wherethey can view their data in various ways (see the DataAnalysis section) or simply quit and leave.Feedback optionsandexperiment execution.Theupper middle sectionofthe GUIs main window providesfeed
40、back options, as well as the experiment executionbutton, entitled Run. Check boxes are utilized here toprovide two feedback options: visual and auditory. Theseboxes are not mutually exclusive; they can be indepen-dently selected and deselected by the user. The effectofclicking on the Run button is to execute the filter ex-periment by calling up MATLAB routines that performthe following operations. First, the requested data set isloaded,andthesavedparameter valuesareevaluated.Thena stimulus-creat
41、ion routine creates and stores the requi-site numberofstimuli needed for the first blockofex-perimental trials (i.e., random four-digit strings are gen-erated and filtered). Counterbalancing for the first blockoftrials is carried out, and finally, the computer screen iscleared, and the observer is prompted to initiate the firsttrial withakeypress.Atthe endofeachblock,the observeris prompted with the option to either continue with a newblockoftrials or quit and return to the GUI.DataAnalysisThe f
42、inal sectionofthe GUIs main window, the bot-tom right, is devoted to data analysis. The static text,Blocks Run: 24, indicates that the selected observer,E.o., has collected 24 blocksofdata for Data Set 2. Edittext boxes allow the experimenter to specify analysisofeither the entire setofblocks (the default) or any validsubsetofblocks. Data (probabilities, standard errors, orboth) can be copied to the clipboard by clicking on theCopy to Clipboard buttons. This is particularly usefulfor transferrin
43、g data to a spreadsheet application, such asExcel, or toa graphing application, such as Kaleidagraph.Plot options.Four plot options are provided on theGUI. The first two are displayed as mutually exclusiveradio buttons: Plot lines and No lines.These allowtheuser to choose whether or not to display the lines on thegraph or remove them, displaying only the data pointsand standard error bars. Below the radio buttons are twoplot options check boxes. The first, when checked, plotsthe data on a log du
44、ration scale, whereas the secondchanges the data from probability(p)to a transform ofprobability, -In(1 -p)(see Loftus, Busey,&Senders,1993). Whenanyoneofthese radio or check boxes is se-lected or deselected, the GUI immediately updates thegraph to reflect the change.Graphicaldisplay.The lowerright cornerofthe mainwindow contains a simple graphofthe data that auto-matically updates itselfwhen new observers and/or datasets are selected for analysis. MATLAB provides a mul-titudeofplot options for
45、use in graphing, including vari-ous colors, symbols, line styles, axes labels, legends, andso forth. Graphs in MATLABcan be as simple or ascom-plex as the user desires.Theory FittingIn our example GUI, we have created a second GUIwindow, the theory window, shown in Figure 2, for alltheory-related operations.Linktotheorywindowandexit button.In the GUIsmain window, the user is provided two options for leav-ing the window: a link to the corresponding theory win-dow,oran exit outofthe GUI altogether
46、. The twobuttonsthat perform these tasks are located within a small framein the upper right portionofthe main window. While theExit button simply closes the GUI, the Theory buttonopens the theory window (described below), while leav-ing open the main windowofthe GUI. This allows theuser to click between the two windows, altering observerand/or data set selection for theory fits.Theory window.Figure 2 shows the theory windowcreated to accompany the main windowofthe filter ex-periment GUI. The cod
47、e called by various objects withinthe window utilizes MATLABs optimization tools to fita specific quantitative theory to the filter experiment data.For interested readers, the perceptual theory used to fitdata from the filter experiment is described in numerouspublications (e.g., Busey&Loftus, 1994; Loftus et al.,1993; Loftus&Ruthruff, 1994). Wewill refer to it as thesensory response/information-acquisition rate (SRIAR)theory.Although knowledgeofthe SRIAR theory is not crit-ical for understandin
48、g the essenceofthis article, we pro-vide a verybriefsynopsisofthe parameters, to aid in un-derstanding the GUIs theory window.The SRIAR theoryhas four parameters. First,n(a unitless positive integer)and r (a positive real number with unitsoftime) are theMATLABANDGRAPHICALUSER INTERFACES295Figure2.Theorywindow of thefilterexperimentgraphicuserinterface(GUI).Thiswin-dow is called by theTheorypushbuttonon themainwindowoftheGUIandutilizesMAT-LABsoptimizationroutinesto fit LoftusandBuseyssensoryrespo
49、nse/informationacqui-sitionratetheoryto thefilterexperimentdata.Note:Thisis ablackandwhitereproductionofamulticoloredGUI.parametersofagammafunction that allow generationofan internalsensory responseto astimulusofa giventemporal shape. There is asensory threshold,e,such thatfurther information processing takes place only when thevalueofthesensoryresponse isgreaterthane(eis aunitlessnumberintimatelyconnectedto stimulus con-trast). A parameter c (with unitsoftime-1)controls the rateat which informa
50、tion is acquired from a given stimulus.For application in the present experiment to the issuesofspatial filtering, two additional parameters are required.296HARLEY AND LOFTUSA unitless parameter is added that reflects the relativeweightofassumed high- versus low-spatial-frequencychannels, and finally, because observers appearto be im-perfect in several respects, a performance asymptote,A,is required.The resulting six free parameter values for the SRIARtheory are listed in edit text boxes along t
51、he top left sideofthe theory window. Push buttons allow the user the op-portunity to paste parameters stored on the clipboard, aswell as to copy parameters and theory predictions ontothe clipboard. To find the best-fitting theory parametersforaset of datapoints, the experimenter clicks on the pushbutton entitled Find Best Fit. This button calls MAT-LAB optimization routines that search for the valuesofthe free parameters that minimize errorbetween the dataand thetheory.Once found, the newparamet
52、ers are copiedinto the parameter edit text boxes. Each time the param-eters are changed, eithermanually or via the optimizationroutines, the theory predictions are graphed in the axesas solid lines overlaying the data points. The action initi-ated by the Full Theory button is to smooth out the the-ory curves by calculating and plotting the theory predic-tions over a large numberofexposure durations, ratherthan only at the six experimentally tested durations. Fi-nally,the Reverse Mask checkbox wa
53、s created to allowus to see whether the SRIAR theory could fit the data ifit was misled about whether or not the stimuli in a partic-ular data set were masked (incidentally, we found that itcould not).SUMMARY ANDCONCLUSIONSMATLAB is a convenient platform for the develop-ment and managementofpsychological experiments,owing to its easy-to-use programming language, sophis-ticated graphics features, and statistics and optimizationtools. Through the additionalimplementationoftheBrainard-Pelli Psychop
54、hysics Toolbox, MATLABs in-terpreted programming language gains the power andflexibilityofa low-level language, giving the user closetemporal and spatial control over the CRT display.MATLABs benefits to psychological experimentationcan be further increased through the useofone graphicalfeature in particular, the GUI. A well-designed GUI is apowerful tool for organizing and controlling all aspectsofrunning a psychological experiment, including design,data collection, data analysis, and theory fit
55、ting.REFERENCESBRAINARD,D. H. (1997). The Psychophysics Toolbox.Spatial Vision,10,433-436.BUSEY,T.A.,&LOFTus, G. R. (1994). Sensory and cognitive compo-nentsofvisual information acquisition.Psychological Review, 101,446-469.COHENJ.,MACWHINNEYB.,FLATT,M.,&PROVOST,J. (1993).PsyScope: An interactive graphic system for designing and control-ling experiments in the psychology laboratory using Macintosh com-puters.Behavior Research Methods, Instruments,&Computers, 25,257-271.HANSELMAN,D.,&LITTLEFIELD,
56、B. (1996).Mastering MATLAB: Acomprehensivetutorial and reference.Englewood Cliffs, NJ: Prentice-Hall.LOFTus,G.R.,BUSEY,T.A.,&SENDERS,J.W.(1993). Providing a sen-sory basis for models of visual information acquisition.Perception&Psychophysics,54, 535-554.Lorrus,G.R.,&RUTHRUFF,E. R. (1994). A theoryofvisual informationacquisition and visual memory with special application to intensity-duration tradeoffs.JournalofExperimentalPsychology: Human Per-ception&Performance,20, 33-50.MARCHAND,P. (1999).Gra
57、phics and GUIs with MATLAB(2nd ed.).Boca Raton, FL: CRC Press.OLDS,E. S.,&ENGEL,S. A. (1998). Linearity across spatial frequencyin object recognition.Vision Research,38, 2109-2118.PARISH,D.H.,&SPERLING,G.(1991). Object spatial frequencies, reti-nal spatial frequencies, noise, and the efficiencyofletterdiscrimina-tion.Vision Research,31, 1399-1415.PELLI,D. G.(1997). The VideoToolbox software for visual psycho-physics: Transforming numbers into movies.Spatial Vision,10,437-442.ST.JAMES,J.D.,&SCHNE
58、IDER,W.(1991). Student MEL software sup-portforinstructorsandteachingassistantsin researchmethodscourse.Behavior Research Methods, Instruments,&Computers, 23,149-154.SCHNEIDER,W. (1989). Enhancing a standard experimental deliverysystem (MEL) for advanced psychological experimentation.Behav-ior Research Methods, Instruments,&Computers,21, 240-244.YEE, P.L., &VAUGHANJ. (1999). AWeb-accessibletutorial forPsyScope based on classic experiments in human cognition.Behav-ior Research Methods, Instrument
59、s,&Computers,31, 107-112.NOTESI.The Math Works announced in 1998 that it would no longer sup-port the Macintosh past version 5.2.Itis hoped that The Math Workswill reconsider this decision in viewofthe Macintoshs resurgence.2. Costwise, MATLAB is an excellentalternative to other experimentcreation software packages. The student versionofMATLAB (which isa full version, not scaled down) currently sells for $99, with optionaltoolboxes, such as the Statistics and Optimization toolboxes, costing anad
60、ditional $59 each. The academic costofMATLAB for nonstudents is$500 with optional toolboxes costing an additional $200 each.3. Todownloador to obtain furtherinformationaboutthe freePsychophysics Toolbox available for both Macintosh and Windows, goto http:/color.psych.ucsb.edu/psychtoolboxl4. For use in this paper, the stimulus color listed for observer E.U.inFigureIwas altered from its actual valueof146 to a lower valueof115.This was done to increase the contrastofthe stimuli pictured in the bot-tom leftcomerofthe GUI and, hence, improve the visibilityofthesestimuli for the reader.(Manuscript received NovemberI,1999;revision accepted for publication February 25, 2000.)
《外文翻译原文-MATLAB和图形用户界面进行实验管理的工具》由会员M****1分享，可在线阅读，更多相关《外文翻译原文-MATLAB和图形用户界面进行实验管理的工具》请在金锄头文库上搜索。
