Novelty detector neurons in the mammalian auditory midbrain

No el y de ec o neu ons in he mammalian audi o y midb ain Da id Pé ez-González,1 Manuel S. Malmie ca1 and Ellen Co ey1,2 1 Audi o y Neu ophysiology Uni , Labo a o y o he Neu obiology o Hea ing, Facul y o Medicine, Uni e si y o Salamanca and Ins i u e o Neu oscience o Cas illa y León, Salamanca, Spain 2 Depa men o Psychology, Uni e si y o Washing on, Sea le, WA 98195, USA Keywo ds: a en ion, habi ua ion, in e io colliculus, epe i ion a e, s imulus-speci ic adap a ion, empo al con ex Abs ac No el s imuli in all senso y modali ies a e highly e ec i e in a ac ing and ocusing a en ion. S imulus-speci ic adap a ion (SSA) and b ain ac i i y e oked by no el s imuli ha e been s udied using popula ion measu es such as imaging and e en - ela ed po en ials, bu he e ha e been ew s udies a he single-neu on le el. In his s udy we compa e SSA ac oss di e en popula ions o neu ons in he in e io colliculus (IC) o he a and show ha a subclass o neu ons wi h apid and p onounced SSA espond selec i ely o no el sounds. These neu ons, loca ed in he do sal and ex e nal co ex o he IC, ail o espond o mul iple epe i ions o a sound bu b ie ly eco e hei exci abili y when some s imulus pa ame e is changed. The inding o neu ons ha espond selec i ely o no el s imuli in he mammalian audi o y midb ain sugges s ha hey may con ibu e o a apid subco ical pa hway o di ec ing a en ion and ⁄ o o ien ing esponses o no el sounds. In oduc ion An impo an unc ion o he audi o y sys em is o di e en ia e beha iou ally unin e es ing pa e ns o sound, which a e o en epe i i e, om no el sounds ha may equi e a en ion o ac ion. Neu ons in all pa s o he in e io colliculus (IC) a e known o show dec eased esponsi eness o ains o iden ical s imuli (Palombi & Caspa y, 1996; Nuding e al., 1999), especially when hese a e p esen ed a high epe i ion a es. Howe e , o da e he e ha e been no sys ema ic a emp s o cha ac e ize he ela ion be ween he obse ed esponse dec emen o epe i ions o iden ical s imuli and he IC neu ons’ esponses o no el s imuli. This s udy ocuses on neu ons in he do sal and ex e nal co ical a eas o he a IC ha showed a apid and p onounced dec emen in esponsi eness o ains o iden ical s imuli, i.e. s imulus-speci ic adap a ion (SSA) e en a low epe i ion a es, bu b ie ly eco e ed hei esponsi eness whene e some s imulus pa ame e was changed. As a consequence, hese neu ons esponded selec i ely o no el s imuli. Thei p ope ies a e consis en wi h he ange o s imula ion pa adigms ha p oduce misma ch nega i i y (MMN), an e en - ela ed po en ial associa ed wi h unexpec ed s imuli in humans and animals (Nää änen, 1995). Ma e ials and me hods Su gical p ocedu es Expe imen s we e pe o med on 21 adul male a s wi h body weigh s be ween 200 and 365 g. Su gical anaes hesia was induced and main ained wi h u e hane (1.5 g ⁄ kg, i.p.), wi h supplemen a y doses (0.5 g ⁄ kg, i.p.) gi en as needed. U e hane was chosen as an anaes he ic because i s e ec s on mul iple aspec s o neu al ac i i y including inhibi ion and spon aneous i ing a e known o be less han hose o ba bi u a e anaes he ics (e.g. Ha a & Ha is, 2002; Neue e al., 2004). The achea was cannula ed, and a opine sulpha e (0.05 mg ⁄ kg, s.c.) was adminis e ed o educe b onchial sec e ions. Body empe a u e was main ained a 38 C ± 1 C. De ails o su gical p epa a ion we e as desc ibed elsewhe e (Malmie ca e al., 2003, 2005; He nández e al., 2005). The animal was placed in a s e eo axic ame in which he ea ba s we e eplaced by hollow specula ha accommoda ed a sound deli e y sys em. All expe imen s we e ca ied ou wi h he app o al o , and using me hods con o ming o he s anda ds o , he Uni e si y o Salamanca Animal Ca e Commi ee. Acous ic s imuli and elec ophysiological eco ding A c anio omy was pe o med o expose he ce eb al co ex and he ce ebellum o e lying he IC. A ungs en elec ode (Me ill & Ainswo h, 1972) was lowe ed h ough he co ex and used o eco d ex acellula single uni esponses in he IC. Neu on loca ion in he IC was based on s e eo axic coo dina es, physiological c i e ia o ono opici y and esponse eliabili y (Palombi & Caspa y, 1996; Rees e al., 1997; Nuding e al., 1999; Syka e al., 2000; Malmie ca e al., 2003, 2005) as well as his ological e i ica ion using elec oly ic lesions (5–10 lA o 5–10 s) o ma k he si es whe e no el y esponses we e eco ded as well as si es o be used o econs uc ion o elec ode acks (Rees e al., 1997; Syka e al., 2000; Malmie ca e al., 2003). S imuli we e deli e ed h ough a sealed acous ic sys em (Rees e al., 1997; Malmie ca e al., 2003, 2005) using wo elec os a ic loudspeake s (TDT–EC1) d i en by wo TDT– ED1 modules. Pu e one bu s s, noise bu s s, and equency and ampli ude modula ed s imuli we e gene a ed and deli e ed o one o bo h ea s unde compu e con ol using he TDT Sys em 2 (Tucke -Da is Technologies, Alachua, FL, USA) ha dwa e and cus om so wa e (Fau e e al., 2003). Typically, ones we e 75 ms du a ion and 5 ms ise ⁄ all ime, a he neu on’s cha ac e is ic equency (CF), wi h selec ed pa ame e s a ied one a a ime du ing es ing. The elec ode was ad anced using a Bu leigh mic od i e. Ac ion po en ials we e eco ded wi h a BIOAMP ampli ie (TDT), he 10· ou pu o which was u he ampli ied and bandpass- il e ed (TDT PC1; c, 500 Hz and 3 kHz) be o e passing h ough a spike disc imina o (TDT SD1). Spike imes we e logged on a compu e by eeding he ou pu o he spike disc imina o in o an e en ime (TDT ET1) synch onized o a iming gene a o (TDT TG6). S imulus gene a ion and on-line da a isualiza ion we e con olled wi h cus om so wa e. Spike imes we e displayed as do as e s o de ed by he acous ic pa ame e a ied du ing es ing. Pe is imulus as e g ams we e p oduced wi h Igo P o so wa e (Wa eMe ics, Inc.). Sea ch s imuli we e pu e ones, noise bu s s, and sinusoidally ampli ude-modula ed (SAM) o equencymodula ed (SFM) signals. To he ex en possible, he app oxima e equency uning o he cell was audio isually de e mined a he s imulus du a ion ha e oked he s onges spiking. The minimum h eshold and CF o he cell we e ob ained by an au oma ed p ocedu e wi h 10–25 s imulus epe i ions a each equency and in ensi y s ep. The equency esponse a ea was plo ed using EXCEL and SIGMAPLOT so wa e. S imuli we e p esen ed in h ee di e en modes. In ‘ andom’ p esen a ion mode, alues o he a iable pa ame e we e andomized ac oss ials. In ‘non-in e lea ed’ o ‘block’ mode, a block o 10–25 iden ical s imuli was p esen ed, a e which he pa ame e alue was changed, and ano he block o ials was p esen ed, wi h his p ocedu e being epea ed a each s ep om he s a alue o he end alue o he a iable pa ame e . In ‘sequen ial’ mode, a single s imulus was p esen ed a each pa ame e alue, s epwise om he s a alue o he end alue, wi h he sequence epea ed un il he desi ed numbe o epe i ions was ob ained. F equency was usually a ied in 10 loga i hmic s eps spanning he es ima ed equency esponse a ea o he neu on, a 10 o 20 dB abo e i s es ima ed h eshold. Ampli ude was usually a ied in 5 o 10 dB s eps, a he es ima ed CF. In block o sequen ial mode he usual sequence was om a lowe a enua ion (i.e. highe ampli ude) o a highe a enua ion (lowe ampli ude). Du a ion was usually a ied in en linea s eps om 2 ms o 200 ms o a one a CF, 10 o 20 dB abo e h eshold. The modula ion a e o SAM and SFM s imuli was a ied in en loga i hmic s eps om 20 o 2000 Hz. To quan i y a neu on’s endency o espond o no el s imuli, o each se o s imulus condi ions we calcula ed a ‘no el y esponse index’ (NRI) using he ollowing o mula: NRI ¼ (P1 ) P2…n) ⁄ Pmax Whe e P1 is he p obabili y o a esponse o he i s p esen a ion o a s imulus, P2…n is he p obabili y o a esponse o subsequen p esen a ions o he same s imulus, and Pmax is he la ge o he wo alues. Using his o mula, he no el y esponse index would be +1 i he neu on esponded only on he i s ial and no o he , i would be )1 i i did no espond on he i s ial, bu esponded on o he s, and 0 i he p obabili y o esponding was equal on he i s ial and subsequen ials. His ological e i ica ion o eco ding si es A he end o each expe imen he animal was gi en a le hal dose o sodium pen oba bi al and pe used ansca dially wi h phospha e bu e ed saline (0.5% NaNO3 in PBS) ollowed by ixa i e (a mix u e o 1% pa a o maldehyde and 1% glu a aldehyde in a Ringe ’s solu ion). Sagi al o ans e se sec ions (50 lm) we e cu on a eezing mic o ome. E e y o he sec ion was s ained wi h 0.1% hionin blue o acili a e iden i ica ion o cy oa chi ec u al bounda ies. Resul s We eco ded om 409 single neu ons h oughou he IC while p esen ing mul iple epe i ions o a sound (see Ma e ials and me hods o de ails). S imulus-speci ic adap a ion (SSA) was de ined as a esponse dec emen o 50% o mo e o e he cou se o en iden ical s imulus p esen a ions. A small subpopula ion o neu ons exhibi ed a high deg ee o SSA a ela i ely slow epe i ion a es, bu b ie ly and eliably eco e ed hei esponsi eness whene e some s imulus pa ame e was changed. As a consequence, hese neu ons esponded selec i ely o no el s imuli. Figu e 1A shows an example o such a neu on. In his case, blocks o 100 epe i ions o a gi en equency we e p esen ed in ascending o de s a ing wi h a block o ials a 7.0 kHz and ending wi h a block a 30.5 kHz. This cell could be d i en a equencies spanning a la ge pa o he a ’s audi o y ange, bu only o a ew p esen a ions a any gi en equency. Thus, ins ead o being selec i e o a speci ic equency ange, he neu on was selec i e o any equency ha had no ecen ly been p esen ed. The ime be ween he las s imulus o one block and he i s s imulus o he nex block was he same as he ime be ween s imuli wi hin a block, so he ansi ion om one block o he nex was compa able o he ‘oddball’ pa adigm ha has been used in o he expe imen s (e.g. Ulano sky e al., 2003). Fo compa ison, Fig. 1B shows da a om a neu on ha did no show SSA o mul iple s imulus p esen a ions a a cons an equency, no enhanced esponses o no el equencies. Al hough many neu ons in he IC, including hose in he cen al nucleus, exhibi ed some deg ee o SSA in esponse o a epea ed s imulus, he majo i y equi ed epe i ion a es highe han 4 ⁄ s o elici i . A small subpopula ion o neu ons showed SSA a all o he epe i ion a es es ed, down o 0.5 ⁄ s, and esponded in a highly selec i e manne o no el s imuli. Neu ons ha showed SSA a a epe i ion a e o 4 ⁄ s o less, wi h a e u n o he o iginal le el o esponse o one o mo e ials when a s imulus pa ame e was changed we e classi ied as ‘no el y’ uni s. Figu e 2 shows an example o a no el y neu on’s esponse a h ee di e en epe i ion a es. A each epe i ion a e, blocks o 25 iden ical s imuli we e p esen ed in ascending o de o equency (24.5 kHz o 30.2 kHz). The epe i ion a e was hen changed om high (4 ⁄ s) o low (1 ⁄ s). A a a e o 4 ⁄ s, habi ua ion was such ha he neu on esponded only o he i s s imulus p esen a ion o he en i e sequence. A slowe epe i ion a es (2 ⁄ s and 1 ⁄ s), he p obabili y o esponse o he i s p esen a ion o a no el equency was 100%, bu he p obabili y o a esponse o subsequen p esen a ions was s ill ex emely low, less han 10%. I is in e es ing o no e ha he esponse la ency o his neu on and he one illus a ed in Fig. 1A was always sho e in esponse o he i s ial o a block han i was o any o he subsequen ials in ha block o which he neu on esponded. Using he abo e c i e ia, 25 uni s showed no el y selec i i y. Th ee o hese uni s we e no his ologically localized, bu o he emaining 22, 19 we e loca ed in he ex e nal co ex o he IC and h ee in he do sal co ex o he IC (Malmie ca e al., 1993; Table 1). None o hese uni s was loca ed in he cen al nucleus o he IC. The pe cen age o no el y uni s was app oxima ely 6% o all IC uni s eco ded, and 16.5% o hose his ologically localized ou side he cen al nucleus. In cases whe e we eco ded mul iuni ac i i y in he icini y o no el y uni s, i ypically exhibi ed apid SSA and no el y esponses simila o hose o single uni s, sugges ing ha no el y uni s a e loca ed in clus e s. Al hough he expe imen s we e pe o med in anaes he ized animals, i seems unlikely ha he high deg ee o habi ua ion and no el y esponses we e due o he anaes hesia as simila no el y esponses ha e been obse ed in he IC o awake ba s du ing he cou se o o he s udies (Casseday & Co ey, 1996; E. Co ey, unpublished obse a ions). Figu e 3 illus a es he ela ion be ween SSA and epe i ion a e o he cells ha we classi ied as no el y uni s compa ed o he popula ion o IC uni s as a whole. The no el y esponse index, calcula ed as desc ibed in he Ma e ials and me hods, is a measu e o he neu on’s p obabili y o a esponse o he i s p esen a ion o a speci ic se o pa ame e alues compa ed o esponses o all o he ials unde ha same se o pa ame e alues. The no el y esponse indices o mos IC neu ons (le column) a e dis ibu ed a ound ze o a all epe i ion a es, while hose o neu ons classi ied as no el y uni s a e skewed owa d posi i e alues a all epe i ion a es. No e ha a he as es a es (> 5 ⁄ s), some IC uni s no classi ied as no el y uni s had indices o +1. This sugges s ha some uni s in he cen al nucleus o he IC show SSA a high epe i ion a es, and ha hei esponses, like hose o he no el y uni s, can be es o ed by changing a s imulus pa ame e . The no el y uni in Fig. 1A, like o he simila uni s om which we eco ded, had i ually no spon aneous ac i i y, so i s i ing eliably signalled he occu ence o a change in he s imulus. The ac ha no el y uni s ypically could be d i en ac oss a b oad ange o s imulus pa ame e alues sugges s ha hey ecei e highly con e gen inpu . Pa ame e s ha , when changed, empo a ily es o ed esponses o no el y uni s included equency (in 87% o cases whe e his pa ame e was a ied), in ensi y (71%), du a ion (31%), he modula ion a e o an SAM s imulus (70%) o he modula ion a e o an SFM s imulus (67%). SSA and es o a ion o esponse by a change in he s imulus occu ed e en hough he epe i i e and no el s imuli we e sepa a ed by iden ical in e s imulus in e als, indica ing ha habi ua ion was s imulus-speci ic and no due solely o in insic p ope ies o he neu on ha se i s eco e y ime a e i ing. To ule ou he possibili y ha no el y esponses we e d i en by a speci ic di ec ion o change in s imulus ene gy, a subse o eigh neu ons we e es ed wi h bo h inc easing and dec easing se ies o alues o a enua ion and ⁄ o du a ion in sequen ial and block modes. The e was no sys ema ic co ela ion be ween magni ude o no el y esponse index and di ec ion o change in ei he o hese p esen a ion modes. All o he no el y uni s om which we eco ded we e ansien esponde s. This is consis en wi h he inding ha a la ge p opo ion o uni s in he do sal pe iphe al egions o he mouse IC ha e ansien onse esponses (Ree z & Eh e , 1999). No el y uni s we e dis ibu ed h oughou he en i e audible equency ange o he a (He ne e al., 1994), and had h esholds simila o hose o he o he IC neu ons om which we eco ded. Because no el y uni s ypically showed a comple e cessa ion o i ing a e one o a ew p esen a ions o any gi en s imulus, i was o en di icul o de e mine hei equency esponse a eas and o he con en ional o ms o uning o audi o y s imulus pa ame e s. One would expec neu ons ha exhibi SSA o espond be e o andomized s imuli han o blocks o iden ical s imuli, so we es ed bo h no el y neu ons and hose ha did no show SSA using andomized s imuli, blocks o iden ical s imuli, and sequen ial a ia ion o a s imulus pa ame e om a beginning alue o an ending alue, wi h he sequence epea ed en imes. Figu e 4A and B compa e esponse as a unc ion o sound ampli ude, collec ed unde he h ee di e en s imulus p esen a ion pa adigms o a no el y neu on and one ha did no exhibi SSA. Fo he neu on ha did no show SSA, he h ee cu es a e supe imposed. The no el y uni only esponded o a ew p esen a ions a any gi en ampli ude, so i s esponses o iden ical s imuli p esen ed in blocks we e g ea ly educed compa ed o he andom condi ion. When s imuli we e p esen ed in sequen ial mode, he esponse o he beginning alue equalled ha o andom s imula ion, because i always ep esen ed a la ge change ela i e o he ending alue. Howe e , esponses o o he alues we e educed compa ed o he andom condi ion. This gene al pa e n was seen in all no el y uni s es ed. Fo no el y neu ons, he highes spike coun s we e always ob ained in andom p esen a ion mode and he lowes in block mode. Spike coun s e oked by sequen ial p esen a ion a ied om neu on o neu on, bu we e gene ally highe han hose ob ained in block mode. This pa e n o esponse o di e en modes o s imulus p esen a ion was simila ega dless o wha pa ame e was a ied. Figu e 4C compa es he equency esponse a eas o a no el y neu on and one ha did no show SSA, measu ed using he h ee s imulus p esen a ion modes. Again, he esponse o he no el y uni was g ea ly educed in block mode compa ed o andom o sequen ial mode, bu he esponse o he non-habi ua ing neu on was simila in all h ee modes. Figu e 5 shows popula ion da a on he neu ons’ no el y esponse index in he h ee p esen a ion modes. The alues o he gene al popula ion o IC neu ons (le column) we e dis ibu ed a ound ze o o all h ee modes, while hose o no el y uni s ( igh column) we e clea ly skewed owa ds posi i e alues e en in sequen ial and andom p esen a ion modes. This inding sugges s ha no el y uni s we e sensi i e no only o changes in simple pa ame e s such as equency o ampli ude, bu also o changes in mo e complex sound pa e ns. Figu e 6A and B compa es he esponses o a no el y uni and one ha did no show SSA (non-habi ua ing) o a sinusoidally ampli ude-modula ed (SAM) s imulus a di e en modula ion a es. The esponses o he non-habi ua ing neu on we e e y simila in all h ee p esen a ion modes. The no el y neu on esponded o SAM a i s bes modula ion a e on nea ly e e y ial in andom mode, bu in block mode i only esponded on he i s ial o a block. In sequen ial mode, esponses we e also poo , indica ing ha he p og essi e changes in modula ion a e in his mode we e no as e ec i e in d i ing he cell’s esponse as we e he unp edic able changes in andom mode. E en when p esen ed wi h s imuli ha con ained epe i i e pa e ns o change such as sinusoidal ampli ude modula ion (SAM), no el y uni s esponded a he beginning o he s imulus, bu ne e showed ongoing phase locking. The inding ha no el y neu ons did no phaselock o SAM is consis en wi h he idea ha he a e o modula ion wi hin a s imulus was su icien ly high o cause supp ession o esponses a e he i s ew modula ion cycles, and ha he highly p edic able pa e n o change wi hin each s imulus was no su icien o e e se he habi ua ion. The ac ha hese ansien ly esponding neu ons we e o en uned o a speci ic modula ion equency can be explained i we assume ha hey expe ience in e ac ing pa e ns o exci a ion and inhibi ion elici ed by he i s wo cycles o he s imulus, wi h acili a ion a some pe iods and inhibi ion a o he s (Co ey e al., 1996; Casseday e al., 1997). No el y uni s hus appea o habi ua e o a complex s imulus pa e n, indica ing ha changes o ampli ude alone a e no su icien o d i e hem, i he changes ollow a apid and p edic able sequence. Ra he , wha is needed is a change om an ‘expec ed’ pa e n o change o an ‘unexpec ed’ one. Such a esponse sugges s ha neu ons in he do sal and ex e nal co ex o he IC possess ela i ely complex compu a ional capaci ies compa ed o neu ons a lowe le els, and implies a so o ‘p imi i e in elligence’ (Nää änen e al., 2001). Fig. 2. Responses o a no el y neu on a h ee di e en s imulus epe i ion a es. The do as e s show esponses o ou di e en equencies, p esen ed in blocks o 25 iden ical s imuli each. A each epe i ion a e, blocks we e p esen ed in ascending o de o equency (24.5 kHz o 30.2 kHz). Repe i ion a e was changed om high (4 ⁄ s) o low (1 ⁄ s). A ows indica e spikes i ed in esponse o he i s p esen a ion o each equency block. The in e al be ween he las p esen a ion o one block and he i s p esen a ion o he new block was he same as ha be ween p esen a ions wi hin a block. Thus, he i s p esen a ion o a new equency ep esen s a no el, o ‘oddball’ s imulus. The inse s ep esen he p obabili y o a esponse occu ing on he i s ial a a gi en equency (N, g ey ba s) s. subsequen p esen a ions o ha equency (R, black ba s). Fig. 3. Dis ibu ion o no el y esponse indices o h ee di e en epe i ion a e anges, wi h equency as he a iable pa ame e . The no el y esponse index was calcula ed o each block o ials unde a gi en se o condi ions, so n ep esen s numbe o blocks o which da a we e ob ained ac oss all no el y uni s ( igh column) and all o he IC uni s (le column). Heigh s o ba s a e no malized as pe cen age o al. Signi ican di e ences we e ound be ween no el y uni s and o he uni s a epe i ion a e >5 ⁄ s (P ¼ 0.021), and epe i ion a e 2.5–5 ⁄ s (P < 0.001). S a is ically signi ican di e ences we e also ound o ‘o he IC uni s’ be ween epe i ion a es >5 ⁄ s and 2.5–5 ⁄ s (P ¼ 0.013) and epe i ion a es o 2.5–5 ⁄ s and <2.5 ⁄ s (P ¼ 0.049). Fo no el y uni s he e was also a signi ican di e ence be ween epe i ion a es >5 ⁄ s and <2.5 ⁄ s (P ¼ 0.049) and epe i ion a es o 2.5–5 ⁄ s and <2.5 ⁄ s (P ¼ 0.027). Fig. 4. E ec o s imulus p esen a ion mode on a e-le el unc ions and equency esponse a eas. Ra e-le el unc ions o a no el y neu on (A) and an IC neu on ha did no exhibi SSA (non-habi ua ing) (B) measu ed using h ee di e en s imulus p esen a ion pa adigms. In bo h plo s, ed lines indica e andomized s imulus p esen a ion; g een lines indica e sequen ial p esen a ion; blue lines indica e block p esen a ion. (C) F equency esponse a ea o a no el y neu on measu ed using h ee di e en s imulus p esen a ion pa adigms. (D) F equency esponse a ea o an IC neu on ha did no exhibi SSA (nonhabi ua ing) measu ed using he same h ee s imulus p esen a ion pa adigms. Fig. 5. Dis ibu ion o no el y esponse index alues o all no el y uni s ( igh column) and all o he IC uni s (le column) in h ee di e en s imulus p esen a ion modes. The no el y esponse index was calcula ed o each block o ials unde a gi en se o condi ions, so n ep esen s numbe o blocks o which da a we e ob ained. The g ea es skew owa d posi i e alues was o no el y uni s in block p esen a ion mode, bu he e was a s a is ically signi ican di e ence (P < 0.001) be ween no el y neu ons and o he IC neu ons in all h ee p esen a ion modes. Fig. 6. E ec o s imulus p esen a ion mode on esponse o sinusoidal ampli ude modula ions (SAM). Responses o a no el y uni (A) and a non-habi ua ing uni in he IC (B) o SAM s imuli p esen ed using h ee di e en s imulus pa adigms. Plo s show spike coun s as a unc ion o modula ion equency o andomized, sequen ial and block p esen a ion as well as he andomized pa adigm epea ed again o con ol o e ec s o long- e m s imula ion his o y. All modula ions we e a a dep h o 100%; he ca ie equency was he neu on’s CF. All s imuli had a peak ampli ude o 20 dB abo e h eshold and a du a ion o 100 ms. Fc, ca ie equency. Fig. 7. La ency dis ibu ion o no el y uni s and o he IC uni s. No malized dis ibu ion o a e age i s spike la encies o no el y neu ons (black ba s) and o he IC neu ons (g ey ba s). in insic p ope ies o IC neu ons, o some combina ion o hese and ⁄ o o he ac o s. Tables Table 1. Dis ibu ion o di e en classes o neu ons in he cen al nucleus, do sal co ex and ex e nal co ex o he in e io colliculus (IC) Non-no el y neu ons No el y neu ons (n) (%) (n) (%) Cen al IC 95 100 0 0 Do sal IC 25 89.3 3 10.7 Ex e nal IC 86 81.9 19 18.1 To al 206 90.4 22 9.6 The numbe o neu ons in his da a se is less han he o al sample because i includes only hose ha we e a si es con aining ma ks o his ological localiza ion.