Prefrontal connections of the parabelt auditory cortex in macaque monkeys
Post on 14-Sep-2016
Embed Size (px)
.Brain Research 817 1999 4558
Prefrontal connections of the parabelt auditory cortex in macaque monkeysTroy A. Hackett, Iwona Stepniewska, Jon H. Kaas )
Department of Psychology, Vanderbilt Uniersity, 301 Wilson Hall, 111 21st Aenue South, Nashille, TN 37240, USAAccepted 27 October 1998
In the present study, we determined connections of three newly defined regions of auditory cortex with regions of the frontal lobe, andhow two of these regions in the frontal lobe interconnect and connect to other portions of frontal cortex and the temporal lobe in macaquemonkeys. We conceptualize auditory cortex as including a core of primary areas, a surrounding belt of auditory areas, a lateral parabelt oftwo divisions, and adjoining regions of temporal cortex with parabelt connections. Injections of several different fluorescent tracers and
. . .wheat germ agglutinin conjugated to horseradish peroxidase WGAHRP were placed in caudal CPB and rostral RPB divisions of the .parabelt, and in cortex of the superior temporal gyrus rostral to the parabelt with parabelt connections STGr . Injections were also placed
in two regions of the frontal lobe that were labeled by a parabelt injection in the same case. The results lead to several major conclusions.First, CPB injections label many neurons in dorsal prearcuate cortex in the region of the frontal eye field and neurons in dorsal prefrontalcortex of the principal sulcus, but few or no neurons in orbitofrontal cortex. Fine-grain label in these same regions as a result of aWGAHRP injection suggests that the connections are reciprocal. Second, RPB injections label overlapping prearcuate and principalsulcus locations, as well as more rostral cortex of the principal sulcus, and several locations in orbitofrontal cortex. Third, STGr injectionslabel locations in orbitofrontal cortex, some of which overlap those of RPB injections, but not prearcuate or principal sulcus locations.Fourth, injections in prearcuate and principal sulcus locations labeled by a CPB injection labeled neurons in CPB and RPB, with littleinvolvement of the auditory belt and no involvement of the core. In addition, the results indicated that the two frontal lobe regions aredensely interconnected. They also connect with largely separate regions of the frontal pole and more medial premotor and dorsalprefrontal cortex, but not with the extensive orbitofrontal region which has RPB and STGr connections. The results suggest that both RPBand CPB provide the major auditory connections with the region related to directing eye movements towards stimuli of interest, and thedorsal prefrontal cortex for working memory. Other auditory connections to these regions of the frontal lobe appear to be minor. RPB hasconnections with orbitofrontal cortex, important in psychosocial and emotional functions, while STGr primarily connects with orbital andpolar prefrontal cortex. q 1999 Elsevier Science B.V. All rights reserved.
Keywords: Superior temporal gyrus; Principal sulcus; Prearcuate; Orbitofrontal
In the present study, we show how three higher-orderand recently defined subdivisions of auditory or auditory-related cortex are connected with specific regions of thefrontal lobe in macaque monkeys. The results of a numberof recent anatomical and physiological studies of monkeysw x26,27,32,34,38,39,47,48 indicate that auditory cortex iscomplexly organized with as many as 15 or more corticalfields, comprising four or more distinct levels of process-
w x.ing see Refs. 26,33 . A core of three primary-like inter-connected fields receives dominant inputs from the ventral
) Corresponding author. Fax: q1-615-343-8449; E-mail:firstname.lastname@example.org
division of the medial geniculate complex and distributesto a narrow surrounding belt of cortex that contains ap-proximately eight fields representing a second level ofprocessing. The belt areas are further interconnected with aparabelt region on the exposed surface of the superiortemporal gyrus, adjacent to the belt. Differences in connec-tions of the rostral and caudal halves of the parabelt withthe belt suggest that the parabelt has at least these twosubdivisions. Cortex on the superior temporal gyrus imme-diately rostral and caudal to the parabelt, as well as cortexon the upper bank of the superior temporal sulcus, hasconnections with the parabelt, but not with the belt andcore. Thus, the parabelt can be conceptualized as a thirdlevel of cortical processing, and the superior temporalcortex with parabelt connections as a fourth level.
0006-8993r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. .PII: S0006-8993 98 01182-2
( )T.A. Hackett et al.rBrain Research 817 1999 455846
One important question is how do these proposed subdi-visions of auditory cortex relate to the frontal lobe? Neu-rons in several regions of the frontal lobe respond toauditory stimuli, most notably in the region of the frontaleye field or area 8A, prefrontal cortex of the principal
wsulcus and even orbitofrontal cortex 2,10,12,13,29,40,52,x55,56,59,61,62,65,67 . None of these regions appear to get
w xdirect inputs from the core areas of auditory cortex 39,44 .Instead, some of the projections to frontal cortex originate
w xin the lateral belt 39,49 , and others stem from a largeextent of the cortex of the adjacent superior temporal gyrus
w xand sulcus 37,1416,30,31,37,42,44,49,57 . Our goal in
the present study was to see how recently defined subdivi-sions of the parabelt and more rostral auditory-relatedcortex of the superior temporal gyrus connect with thefrontal lobe. To do this, injections of different distinguish-able tracers were placed by reference to surface features insuperior temporal cortex just lateral to the Sylvian fissure.Injections were considered to be in the caudal parabelt ifthey labeled caudal belt areas, the rostral parabelt if theylabeled rostral parabelt areas, and rostral to the parabelt ifthey labeled parabelt cortex, but failed to label belt cortex.The placement of more than one tracer in a given animalprovided some information on the divergence and conver-
.Fig. 1. Left hemisphere of the macaque monkey brain. A Lateral view of the left hemisphere showing position of the parabelt auditory cortex and location . .of the two prefrontal tracer injections. B Dorsolateral view of the normalized left superior temporal region showing positions of the core black , belt
. .dark gray , parabelt light gray , and the relative locations of injection sites in the dorsal superior temporal gyrus. Symbols indicate location and casenumber.
( )T.A. Hackett et al.rBrain Research 817 1999 4558 47
gence of the frontal lobe connections across and withinthese three auditory regions. Finally, we placed injectionsof different tracers in two frontal lobe regions that werelabeled by an injection of a third tracer in the parabeltcortex. These injections provided information on the natureof temporal lobe connections with auditory-related regionsof the frontal lobe, and the relationship of these regionswith other parts of the frontal lobe. Some of these results
w xhave been briefly presented before 25 .
2. Materials and methods
Details concerning anesthesia, surgery, and proceduresfor making tracer injections have been described in detailin our previous papers on cortical and thalamic connec-
w x.tions based on the same cases see 26,27 . An abbrevi-ated account of some of the procedures is included here.
Fig. 2. Fluorescence photomicrographs of flattened prefrontal cortex case. . .3, 97-33L . A Injection site and tract of Fluoro-Ruby FR in the arcuate
.sulcus. Lateral is to the left, dorsal is up. Scale bars1 mm. B Tworetrogradely labeled cells near the caudal end of the dorsal principalsulcus labeled by the FR injection in A. Rostral is to the left, dorsal is up.Scale bars100 mm.
. .Fig. 3. Left frontal lobe of the macaque monkey case 96-1 . ADorsolateral view of the block containing the left frontal lobe showing
.major sulci. B Surface view of the same block, unfolded, beforeflattening between glass plates. The cut was made along the centralsulcus. AS, arcuate sulcus; Cing, cingulate cortex; Cl; claustrum; CS,central sulcus; MOS, medial orbital sulcus; PS, principal sulcus. Scalebars10 mm.
2.1. Tracer injections
2.1.1. Parabelt and superior temporal gyrusIn each of three macaque monkeys 2 M. mulatta, 1 M.
.nemestrina , injections of four to six anatomical tracerswere made in a rostrocaudal series along the dorsal supe-
.rior temporal gyrus STG within 35 mm of the lateralsulcus in the left hemisphere, corresponding to the parabelt
.and STG rostral to the parabelt Fig. 1 . The tracers usedwere horseradish peroxidase conjugated to wheat germ
.agglutinin WGAHRP and the following fluorescent . .tracers: Fast Blue FB , Diamidino Yellow DY , Fluoro-
. .Ruby FR , Fluoro-Emerald FE , Rhodamine green beads . .GB , and Rhodamine red beads RB . Because the tracersvary in sensitivity, the amounts and solution concentrations
were varied accordingly 0.03 ml at 2% for WGAHRP;0.250.30 ml at 27% for fluorescent tracers; 4 ml at 10%
.for rhodamine beads . In the present study, the transport ofneuronal tracers associated with 10 STG injections wasanalyzed.
( )T.A. Hackett et al.rBrain Research 817 1999 455848
.Fig. 4. Reconstruction of flattened cortical sections of the left prefrontal cortex case 95-10L showing the locations of labeled cells after multiple tracerinjections into the parabelt region of the left superior temporal gyrus. Symbols mark the locations where retrogradely labeled cells were found. Interruptedlines denote edges of sulci. Inset: locations of injection sites are denoted by symbols in a lateral view of the left hemisphere. AS, arcuate sulcus; Cing,cingulate cortex; d, dorsal; DY, Diamidino Yellow; FB, Fast Blue;; LOS, lateral orbital sulcus; MOS, medial orbital sulcus; PS, principal sulcus; v, ventral;vl, ventrolateral; WGAHRP, wheat germ agglutininhorseradish peroxidase. Scale bars5 mm.
( )T.A. Hackett et al.rBrain Research 817 1999 4558 49
2.1.2. Prefrontal cortex .In one case 97-33L , two retrograde tracers were also
.injected into the prefrontal cortex Fig. 1a . A 0.7-mlinjection of FE was made in the dorsal lip of the caudalprincipal sulcus as close to the edge of the sulcus aspossible, given the vasculature. The syringe entered thecortex at a slight dorsal-to-ventral angle and the injectionwas centered at a depth of 2.0 mm to minimize thepossibility of spread from area 46d to area 9. A second,physiologically guided, injection was made into the dor-sorostral limb of the arcuate sulcus in area 8A. Underketamine anesthesia, a 1-MV tungsten microelectrode at-tached to a hydraulic microdrive was advanced into therostral bank of the arcuate sulcus at an angle roughly
parallel to the surface of the sulcus perpendicular to the.gyral surface . The cortex was stimulated with single
400-ms presentations of a 20-mA electrical pulse at depthintervals of 500 mm until downward eye movements wereelicited between 5 and 6 mm. The electrode was removedand a 0.4-ml injection of FR was made at a depth of 6 mmin the same tract using a 1-ml Hamilton syringe attached tothe microdrive. The syringe was maintained in this posi-tion for 10 min following the injection to minimize spreadof the tracer along the electroderneedle tract. The injec-tion site and retrogradely labeled cells are shown in Fig. 2.The heaviest deposit of FR was confined to the deepest 2mm of the injection site. Dense labeling of cells, terminals,and tracer-filled axons radiated from this portion of theinjection site. A relatively small number of cells andterminals along the remaining 4 mm of the tract werelabeled by the tracer.
The postinjection survival period was seven days when .WGAHRP was used two cases and 12 days when only
.fluorescent tracers were injected two cases . At the end ofthe survival period, a lethal dose of pentobarbital wasadministered. Just before cardiac arrest, the animal wasperfused through the heart with warm saline followed bycold 2% paraformaldehyde dissolved in 0.1 M phosphatebuffer. Immediately after perfusion, the brains were re-moved, photographed, and blocked. A prefrontal block is
shown in Fig. 3a. In three cases 95-10L, 96-1L; and.97-33L , the pia mater and white matter were removed
from the blocks containing the prefrontal cortex. The .cortex was unfolded Fig. 3b and flattened between glass
plates. The flattened blocks were immersed in 30% sucrosesolution overnight, then cut parallel to the pial surface at
.40 mm on a freezing microtome. One case 95-52R wascut in the coronal plane at 50 mm. Alternate series of
.sections were processed for: 1 fluorescent microscopy; .2 WGAHRP using a low artifact tetramethyl benzidine . w x . w x .TMB procedure 24 ; 3 cytochrome oxidase 68 ; 4
w x . w xacetylcholinesterase 23 ; and 5 myelinated fibers 22 .
2.3. Data analysis
Cells labeled with fluorescent tracers were plotted underultraviolet illumination with a Leitz microscope coupled to
an XY plotter Bioquant System, R&M Biometrics,.Nashville, TN . Cells labeled with WGAHRP were plot-
ted under dark field illumination with a drawing tube.Photographs and drawings of each section were madenoting architectonic boundaries, the location of blood ves-sels, and the distribution of labeled cells. A compositedrawing was made from adjacent sections processed for
.label four to five sections , cytochrome oxidase, acetyl-cholinesterase, and myelin by aligning common architec-tonic features and blood vessels. Composite images werecomposed of the reconstructions using Canvas v3.5 soft-ware installed on a Power Macintosh 7200r75 computer .Apple Computers, Cupertino, CA . The final drawings
.were analyzed to reveal: 1 the individual connection .patterns of each tracer injection; 2 the connection pat-
.terns of injections at similar or dissimilar locations; and 3consistency across cases. The symbols used in the finaldrawing demarcate the locations of labeled cells, but notthe actual number.
Photographs were made using a 35-mm Olympus C-35Acamera mounted on an Olympus BH-2 microscope. Colorslides of these sections were scanned at 300 dots per inchwith a Polaroid Sprint Scan color scanner and convertedinto grayscale using Adobe Photoshop v4.0 software .Mountain View, CA . The digitized images were adjustedfor brightness and contrast, cropped, and pasted in theframe where text was added using Photoshop 4.0 software.Except for contrast adjustment and cropping, the imageswere not altered in any way.
2.4. Localization of injections in the parabelt
The parabelt cortex was defined by location along thelateral margin of belt cortex which surrounds the core. The
.three major regions core, belt, and parabelt can be reli-ably distinguished on the basis of layer IIIrIV reactivityfor acetylcholinesterase, cytochrome oxidase, myelin, and
w x.parvalbumin see Ref. 26 . In these preparatio...