Binge like ethanol exposure during the e

Tuan D Tran Department of Psychology East Carolina University Greenville, NC 27858 E-mail: trant@ecu.edu Mark E Stanton Department of Psychology University of Delaware 131 Wolf Hall Newark, DE 19716 Charles R Goodlett Department of Psychology Indiana University-Purdue University at Indianapolis Indianapolis, IN 46202 Binge-Like Ethanol Exposure During the Early Postnatal Period Impairs Eyeblink Conditioning at Short and Long CS–US Intervals in Rats ABSTRACT: Binge-like ethanol exposure on postnatal days (PD) 4–9 in rodents causes cerebellar cell loss and impaired acquisition of conditioned responses (CRs) during ‘‘short-delay’’ eyeblink classical conditioning (ECC), using optimal (280– 350 ms) interstimulus intervals (ISIs) We extended those earlier findings by comparing acquisition of delay ECC under two different ISIs From PD to 9, rats were intubated with either 5.25 g/kg of ethanol (2/day), sham intubated, or were not intubated They were then trained either as periadolescents (about PD 35) or as adults (>PD 90) with either the optimal short-delay (280-ms) ISI, a long-delay (880-ms) ISI, or explicitly unpaired CS and US presentations Neonatal binge ethanol treatment significantly impaired acquisition of conditioning at both ages regardless of ISI, and deficits in the acquisition and expression of CRs were comparable across ISIs These deficits are consistent with the previously documented ethanol-induced damage to the cerebellar–brainstem circuit essential for Pavlovian ECC ß 2007 Wiley Periodicals, Inc Dev Psychobiol 49: 589–605, 2007 Keywords: eyeblink conditioning; fetal alcohol syndrome; neonatal alcohol exposure; interstimulus interval; Pavlovian; short-delay; long-delay; rats INTRODUCTION Heavy prenatal exposure to alcohol results in enduring brain damage and neurodevelopmental disorders, evident both in children diagnosed with fetal alcohol syndrome (FAS) (e.g., Hanson, Streissguth, & Smith, 1978; Jones & Smith, 1973; Stratton, Howe, & Battaglia, Received 27 February 2007; Accepted March 2007 Correspondence to: T D Tran Contract grant sponsor: NIAAA Contract grant numbers: AA11945 and AA09838 (to CRG), AA014288 and T32AA07462 (to MES) Published online in Wiley InterScience (www.interscience.wiley.com) DOI 10.1002/dev.20226 ß 2007 Wiley Periodicals, Inc 1996; Streissguth, Barr, Martin, & Herman, 1980), and in children with a confirmed history of heavy prenatal alcohol exposure but lacking the facial dysmorphia needed for the diagnosis of FAS (Hoyme et al., 2005; Streissguth & O’Malley, 2000) This fetal alcohol spectrum disorder (FASD) has no consensus ‘‘FAS neurobehavioral phenotype’’, but neurobehavioral sequelae can include hyperactivity and attention deficits, deficits in motor coordination, lack of regulation of social behavior or poor psychosocial functioning (Olson, Feldman, Streissguth, Sampson, & Bookstein, 1998; Roebuck, Mattson, & Riley, 1999), and deficits in cognition (Olson et al., 1998; Riley, McGee, & Sowell, 2004), mathematical ability (Aronson, Hagberg, & Gillberg, 1997), verbal fluency (Mattson, Riley, Delis, Stern, & Jones, 1996; Sowell et al., 2001), and spatial memory (Aronson et al., 1997; Hamilton, Kodituwakku, 590 Tran, Stanton, and Goodlett Sutherland, & Savage, 2003; Uecker & Nadel, 1998) Structural magnetic resonance imaging (MRI) studies of FASD children have shown thinning or displacement of the corpus callosum (Sowell et al., 2001), reductions in the size of the anterior cerebellar vermis (Sowell et al., 1996) and the basal ganglia (Archibald et al., 2001), and increased density and narrowing of frontal and inferior parietal/perisylvian cortical gray matter regions (Sowell et al., 2002) The links between the types of prenatal alcoholinduced structural changes in brain and the variability in the type and extent of specific deficits in neurobehavioral functioning have not yet been identified Toward that end, the cerebellum provides an optimal brain region to focus studies directed toward determining the relationships between alcohol-induced structural damage and specific functional outcomes The cerebellum is a target of damage in humans with heavy prenatal alcohol exposure (Clarren, 1986; Riley et al., 2004; Sowell et al., 1996), and FASD children show deficits in cerebellar-dependent behaviors including gait (Driscoll, Streissguth, & Riley, 1990), balance (Roebuck, Simmons, Mattson, & Riley, 1998), and coordinated motor performance (Connor, Sampson, Streissguth, Bookstein, & Barr, 2006; Kyllerman et al., 1985) Structural damage to the cerebellum has been effectively modeled by binge-like alcohol exposure in neonatal rodents during the early postnatal ‘‘brain growth spurt’’, a period of rapid brain development that is comparable to a similar stage in humans that begins around fetal week 24 and extends over the third trimester (Bayer, Altman, Russo, & Zhang, 1993; Dobbing & Sands, 1973; West, 1987; Zecevic & Rakic, 1976) This ‘‘third trimester equivalent’’ includes cerebellar Purkinje cell dendritic outgrowth and synaptogenesis and proliferation and migration of cerebellar granule cells (Altman, 1972a,b, 1982) Dose-dependent cerebellar Purkinje cell and granule cell loss is induced by binge-like alcohol exposure on postnatal days (PD) 4–9 (e.g., Bonthius & West, 1991; Goodlett & Johnson, 1997b, 1999; Pierce, Serbus, & Light, 1993), and PD 4–6 appears to be the time of enhanced vulnerability (e.g., Hamre & West, 1993; Pierce, Williams, & Light, 1999) Cell loss in other neuronal populations directly linked by afferent or efferent projections to the cerebellar cortex also occurs, including in the inferior olive (Napper & West, 1995), the deep cerebellar nuclei (Green, Tran, Steinmetz, & Goodlett, 2002), and the interpositus nucleus (Tran, Jackson, Horn, & Goodlett, 2005) Heavy neonatal binge alcohol exposure produces longlasting deficits on tasks of coordinated motor performance known to depend on cerebellar function (e.g., Goodlett & Lundahl, 1996; Klintsova et al., 1998; Meyer, Kotch, & Riley, 1990) However, the neural circuitry and mechanisms of plasticity underlying these complex motor Developmental Psychobiology DOI 10.1002/dev performance tasks are not well understood In contrast, eyeblink classical conditioning (ECC) is a simple, cerebellar-dependent Pavlovian conditioning task that provides one of the most useful means to assess the functional consequences of alcohol-induced damage to the cerebellum (Goodlett, Stanton, & Steinmetz, 2000; Green, 2004) Similar ECC procedures can be applied across species (including humans) and over development (Ivkovich, Eckerman, Krasnegor, & Stanton, 2000; Woodruff-Pak & Steinmetz, 2000a,b), and the essential circuitry and sites of neuroplasticity mediating the learning is relatively well known from experimental studies in rabbits, rats, and mice (e.g., Anderson & Steinmetz, 1994; Chen, Bao, Lockard, Kim, & Thompson, 1996; Freeman, Barone, & Stanton, 1995; McCormick, Steinmetz, & Thompson, 1985; Steinmetz, Rosen, Chapman, Lavond, & Thompson, 1986), including recent in vivo functional neuroimaging studies in rabbits (Miller et al., 2003) We have shown that neonatal binge-like alcohol exposure in rats significantly impairs the acquisition of conditioned eyeblink responses in weanling rats ($24 days old), periadolescent rats (30–35 days old) and in adults, when the interstimulus interval (ISI, the interval between the CS onset and US onset) is short, i.e., 280– 350 ms (Green, Rogers, Goodlett, & Steinmetz, 2000; Green, Tran, et al., 2002; Stanton & Goodlett, 1998; Tran et al., 2005) The acquisition deficits are also significantly correlated with neonatal alcohol-induced cell loss in the deep neurons of the cerebellum (Green, Tran, et al., 2002), a key site of plasticity supporting the learning of CS–US associations in ECC (Kleim et al., 2002; Lavond, Kim, & Thompson, 1993) In normally developing rats, the emergence of reliable short-delay ECC using a 280-ms ISI occurs between postnatal days (PD) 17–24 (Stanton, Freeman, & Skelton, 1992), and rates of acquisition in 24-day-old rats [measured as percentage of eyeblink conditioned responses (CRs)] is faster using the 280-ms ISI than using ISIs of 560, 1120, or 1190 ms (Freeman, Spencer, Skelton, & Stanton, 1993) The slower conditioning imparted by longer delay intervals in 24-day-old rats has been recently confirmed and also extended to 30-day-old (periadolescent) rats (Claflin, Garrett, & Buffington, 2005); that study also showed that regardless of ISI, the 30-day-old rats conditioned better than 24-day-olds The evidence that short ISIs are optimal for ECC in rats confirms earlier finding in rabbits (Schneiderman & Gormezano, 1966) In contrast, human infants acquire ECC better at intermediate ISIs (650 ms) than either short (250 ms) or long (1250 ms) delays (Claflin & Stanton, 2002; Little, Lipsitt, & Rovee-Collier, 1984) Similarly, aged adult humans condition better with longer ISIs (Woodruff-Pak, Jaeger, Gorman, & Wesnes, 1999) Developmental Psychobiology DOI 10.1002/dev The impairments in short-delay ECC induced by binge neonatal alcohol exposure in rats may reflect a diminished ability to perform the conditioned eyeblink responses at the short-delay ISI used in our initial studies As with infant or aged humans, it is possible that alcohol-exposed rats may require longer delay intervals to perform the CR needed to express the learning Alternatively, longer delay ISIs may produce even more severe ECC impairments in ethanol-exposed rats than for the short-delay ISIs, since the increased demands to control the precise timing of the conditioned response at a non-optimal interval is typically associated with the slower acquisition in intact rats In the following two experiments, we evaluated the latter alternative by comparing the effects of binge neonatal alcohol exposure on ECC using two different interstimulus intervals—the optimal short-delay ISI used in our previous studies (280-ms), and a long-delay ISI (880-ms) used in previous studies of the development of ECC in rats (Claflin et al., 2005; Ivkovich, Paczkowski, & Stanton, 2000) Experiment tested periadolescent rats (around 35 days old at the start of training), whereas Experiment tested adults EXPERIMENT 1: PERIADOLESCENTS The effects of binge neonatal ethanol exposure on shortdelay (280-ms ISI) and long-delay (880-ms ISI) ECC were tested first in periadolescent rats, an age during normal rat development at which delay ECC acquisition and performance has recently emerged (Claflin et al., 2005; Stanton et al., 1992) Long-delay training typically is more difficult for rats to acquire compared to shortdelay training, in part because of the greater demands for proper timing of the CR This experiment tested the hypothesis that the rats given binge neonatal alcohol exposure would exhibit greater impairments in acquiring well-timed, ‘‘adaptive’’ CRs in the long-delay compared to the short-delay procedure Method Subjects and Neonatal Ethanol Treatment Long-Evans rats (m ¼ 45, f ¼ 42) were produced from breeders obtained from Simonsen Laboratories (Gilroy, CA) Breeders were mated overnight in the vivarium of the Department of Psychology at Indiana University-Purdue University at Indianapolis (IUPUI) Vaginal smears were examined early the next morning and females with sperm-positive smears were designated as being on gestational day (GD) of pregnancy On PD 4, litters were randomly assigned to receive intubation treatments or to the unintubated control condition, and culled to eight pups per litter (four males and four females when possible) On PD 4, rats within each litter designated for neonatal intubations on PD 4–9 were randomly assigned within sex to Ethanol-Induced Eyeblink Conditioning Deficits 591 either the ethanol-intubation (EtOH) group, which received two daily intubations of milk ỵ EtOH totaling 5.25 g/kg/day, (2.625 g/kg per intubation; intubations separated by hr), or to the Sham-Intubated (SI) control group, which only received the intubation procedure without any fluids delivered For the intubated litters, the pups were randomly assigned by sex within litter either to the EtOH group or to the SI group with two males and two females per litter in each treatment group The EtOH pups were intubated on PD 4–9 with ethanol in a custom milk formula solution (11.9%, v/v) via intragastric intubation, a procedure that has been described previously (Goodlett, Peterson, Lundahl, & Pearlman, 1997a; Goodlett, Pearlman, & Lundahl, 1998) The EtOH pups were given additional intubations of milk solution (no ethanol), to provide additional calories during the period of peak intoxication; the number of additional intubations needed to provide these additional calories was determined from previous work in our lab and others (Goodlett et al., 1997a, 1998; West, Hamre, & Pierce, 1984) The pups of the litters assigned to the Unintubated Control condition were reared normally as suckle controls (without any intubations), and were weighed daily during the PD 4–9 period All pups were weighed each day during PD 4–9, at weaning (PD 21), and at PD 30 Two hours after the second EtOH intubation on PD 4, a 20-ml blood sample was collected in a heparinized capillary tube from a tail-clip of each intubated pup (EtOH and SI) The tubes were centrifuged and plasma was separated and frozen at À70 C Blood ethanol concentrations (BECs) for the EtOH rats were determined using an oximetric procedure where an Analox GL5 Analyzer (Analox Instruments USA, Inc., Lunenburg, MA) was used to measure the rate of oxygen consumption resulting from oxidation of EtOH in the sample All procedures were approved by the IUPUI Institutional Animal Care and Use Committee Surgery Surgeries were conducted between PD 28 and 36 using procedures described previously (Stanton & Goodlett, 1998; Tran et al., 2005) Rats were anesthetized with Isoflurane gas (Abbott Laboratories, Abbott Park, IL) Differential electromygraphic (EMG) recording wires (Sigmund Cohn, Mt Vernon, NY) were implanted on the upper eyelid muscle of the left eye and a ground wire was placed subcutaneously, posterior to Lambda A bipolar stimulating electrode (Plastics One, Inc., Roanoke, VA) was placed subcutaneously on the periorbital muscle immediately caudal to the left eye for delivery of the US This ‘‘headstage’’ containing EMG recording wires and the plug-end of the bipolar electrode was secured using dental acrylic (Plastics One, Inc.) Each rat was monitored and kept warm during recovery and returned to its home cage Apparatus The apparatus was essentially the same one used by other developmental investigators of ECC (Claflin et al., 2005; Kleim et al., 2002; Stanton & Freeman, 1994) and has been described previously (e.g., Tran et al., 2005) Animals were allowed to move freely in a modified test box constructed with aluminum and clear polycarbonate walls (Med Associates, St Albans, VT) The test box was contained within a soundattenuated chamber (BRS-LVE, Inc., Laurel, MD) The animals’ headstages were connected to wires attached to a commutator 592 Tran, Stanton, and Goodlett Developmental Psychobiology DOI 10.1002/dev Table Median Age in Days (Range) at the Start of Eyeblink Training and Number of Subjects as a Function of Experiment Neonatal Treatment Group Ethanol Intubated (EtOH) Experiment 1: periadolescents Short-delay Paired 34 (R ¼ 30–37), f ¼ 4, m ¼ Explicitly unpaired 37 (R ¼ 36–38), f ¼ 4, m ¼ Long-delay Paired 35 (R ¼ 33–38), f ¼ 5, m ¼ Explicitly unpaired 36 (R ¼ 35–38), f ¼ 3, m ¼ Experiment 2: adults Short-delay Paired 145 (R ¼ 97–202), f ¼ 5, m ¼ Explicitly unpaired 152 (R ¼ 152–202), f ¼ 2, m ¼ Long-delay Paired 151 (R ¼ 92–159), f ¼ 6, m ¼ Explicitly unpaired 151 (R ¼ 102–179), f ¼ 4, m ¼ (Litton Systems, Blacksburg, VA) that was secured on the top chamber Each chamber was equipped with a fan (55–65 dB background noise), house light (15 W), and two piezoelectric speakers, one of which was used for presentation of the tone CS Both the fan and house light were left running during training The US was produced by a constant-current, 60 Hz square wave stimulus isolator (World Precision Instruments, Sarasota, FL) Custom-built ECC system was used to control the delivery of stimuli and recorded the rectified and integrated EMG activity (JSA Designs, Raleigh, NC) Eyeblink Classical Conditioning Procedures Beginning between PD 30 and 38, rats were randomly assigned to one of four training conditions: short-delay paired CS–US training using a 280-ms ISI (PRD280); long-delay paired CS–US training using an 880-ms ISI (PRD880); explicitly unpaired training control using 380 ms tones (UNPRD380); or, explicitly unpaired training control using 980 ms tones (UNPRD980) Not more than one male and one female rat per litter per neonatal treatment condition was assigned to a given training condition Table shows the median age (postnatal days) at the start of eyeblink training, the age range (R) of all groups in both ISI training conditions, and the corresponding sample sizes Rats received six sessions of short- or long-delay eyeblink training Sessions occurred twice a day over three consecutive days; each session within the day was separated by hr Paired CS–US acquisition trials were presented using parameters of Ivkovich, Paczkowski, et al (2000) for short-delay or long-delay conditioning (see Fig 1) Rats in the PRD280 group were exposed to a 380-ms, 2.8-kHz, 80-dB tone CS that overlapped and co-terminated with a 100-ms periocular shock US, to produce an ISI of 280 ms between CS and US onset For rats in the PRD880 group, a longer tone CS (980 ms) overlapped and Sham Intubated (SI) Unintubated Control (UC) 34 (R ¼ 32–36), f ¼ 5, m ¼ 37 (R ¼ 36–38), f ¼ 3, m ¼ 34 (R ¼ 32–36), f ¼ 4, m ¼ 37 (R ¼ 36–38), f ¼ 2, m ¼ 35 (R ¼ 33–38), f ¼ 5, m ¼ 36 (R ¼ 35–38), f ¼ 2, m ¼ 37 (R ¼ 34–38), f ¼ 3, m ¼ 36 (R ¼ 36–38), f ¼ 2, m ¼ 181 (R ¼ 97–202), f ¼ 6, m ¼ 179 (R ¼ 102–202), f ¼ 3, m ¼ 167 (R ¼ 145–187), f ¼ 5, m ¼ 166 (R ¼ 151–186), f ¼ 2, m ¼ 129 (R ¼ 92–179), f ¼ 6, m ¼ 151 (R ¼ 123–179), f ¼ 3, m ¼ 154 (R ¼ 92–167), f ¼ 5, m ¼ 139 (R ¼ 127–157), f ¼ 3, m ¼ co-terminated with the 100-ms shock US, to produce an ISI of 880 ms Paired CS–US sessions for both the short- and longdelay conditioning procedures consisted of 10 blocks of trials with each block including nine paired trials followed by one CSalone trial (100 trials total), with an intertrial interval (ITI) that averaged 30 s Rats that underwent unpaired training received FIGURE Illustration of standard short-delay (D280) and long-delay (D880) eyeblink conditioning procedures In both procedures, the tone CS starts 280 ms into the trial epoch, overlaps, and co-terminates with a shock US that lasts 100 ms The CS duration in D280 is 380 ms, whereas in D880 it is 980 ms The interstimulus interval (ISI) between the CS and US is 280 ms and 880 ms, respectively Developmental Psychobiology DOI 10.1002/dev Ethanol-Induced Eyeblink Conditioning Deficits 593 Table Mean Maximum US Intensities (Average of Six Sessions) and UR Amplitudes (Final 80 Trials [Paired] or Final 80 USOnly Trials [Explicitly Unpaired] of Session 1) as Functions of Short-Delay and Long-Delay Conditioning Paired Traininga Neonatal Treatment Experiment 1: periadolescents Short-delay (D280) EtOH SI UC Long-delay (D880) EtOH SI UC Experiment 2: adults Short-delay (D280) EtOH SI UC Long-delay (D880) EtOH SI UC Explicitly Unpaired Trainingb Maximum US Intensity (mA) UR Amplitude (V) Maximum US Intensity (mA) UR Amplitude (V) 2.6 Ỉ 0.2 2.5 Ỉ 0.2 2.2 Ỉ 0.3 2.7 Ỉ 0.2 2.4 Ỉ 0.2 2.7 Ỉ 0.3 3.42 Ỉ 0.28 2.60 Æ 0.12 4.45 Æ 0.32 4.36 Æ 0.28 4.17 Æ 0.28 3.92 Ỉ 0.29 2.1 Ỉ 0.3 2.1 Ỉ 0.3 1.9 Ỉ 0.3 2.0 Ỉ 0.3 1.7 Ỉ 0.3 1.9 Æ 0.3 4.68 Æ 0.79 4.44 Æ 0.87 4.42 Æ 0.87 5.24 Ỉ 0.87 5.84 Ỉ 0.87 5.68 Ỉ 0.87 1.76 Ỉ 0.3 1.80 Ỉ 0.3 1.43 Ỉ 0.3 1.82 Æ 0.3 1.68 Æ 0.3 1.56 Æ 0.3 2.75 Æ 0.16 2.73 Ỉ 0.19 2.49 Ỉ 0.17 2.03 Ỉ 0.14 2.49 Ỉ 0.12 2.32 Ỉ 0.12 1.9 Ỉ 0.3 2.0 Æ 0.3 1.7 Æ 0.3 1.5 Æ 0.3 1.5 Æ 0.3 1.5 Ỉ 0.3 3.20 Ỉ 0.58 2.68 Ỉ 0.49 2.98 Ỉ 0.58 2.84 Ỉ 0.53 2.62 Ỉ 0.58 2.74 Æ 0.53 Note Data are expressed as means Æ SE a Mean maximum US intensities and UR amplitudes were not significantly different among treatment groups given paired CS–US training b Mean maximum US intensities and UR amplitudes were not significantly different among treatment groups given explicitly unpaired training CS and US presentations that were explicitly not paired together in a trial Rats in the UNPRD380 group were exposed to the same 380-ms tone CS that was used in the PRD280 group, while rats in group UNPRD980 were exposed to the longer 980-ms tone CS that was used in the PRD880 group Explicitly unpaired sessions consisted of 200 trials (100 CS-alone, 90 US-alone, 10 trials programmed without either the CS or US) The trials were presented in a pseudorandom order such that no more than three consecutive presentations of either stimulus occurred The average ITI was 15 s This insured that the number and relative temporal distribution of CS and US presentations was the same in the paired and unpaired groups The US intensity used for conditioning was determined on an individual rat basis, such that at the start of the first session, the shock intensity was initially set at 0.4 mA and increased in 0.2 mA increments over the first 20 trials until the shock evoked consistent eyeblink unconditioned responses (URs) that produced post-stimulation EMG amplitudes that were at or above V (see below) No further adjustments in shock intensity were performed in Session after the 20th trial Subsequent sessions used the US intensity level established in Session If, after the first session, there was one other session in which the UR amplitudes fell below criteria at the beginning of the session, one additional adjustment of the US intensity was completed [up to a maximum of mA, the highest intensity used in the previous parametric ECC study in weanling rats; Freeman et al., 1993] If after that the UR remained low or degraded on a later session, the animal was dropped from the study Ninety-one rats were initially obtained for Experiment 1; four rats were dropped (PRD880: EtOH ¼ 1, SI ¼ 1; UNPRD380: UC ¼ 2) because of unreliable URs, typically due to sub-optimal placement of the bipolar electrodes or loss of signals from the EMG electrodes The mean maximum US intensities during Session were initially analyzed using (Sex)  (ISI)  (Treatment Group) ANOVAs within either the paired or explicitly unpaired training conditions; there were no significant main or interactive effects The means are shown in see Table Eyelid EMG activity was amplified (5,000Â) and bandpass filtered at 500–5,000 Hz with a 12 dB per octave rolloff by a differential ac pre-amplifier, and then rectified and integrated by a dc integrator (10Â) before being passed to a computer for storage, in volt (V) units The integration time constant of the integrator was 20 ms and the overall voltage gain from the preamplifier was 50,000 For Group PRD280, EMG signals were sampled in 2.5 ms bins during each 800 ms trial epoch and for Group PRD880 signals were sampled in 3.5 ms bins during each 1,400 ms trial epoch Each trial epoch was divided into five time periods (see Fig 1): (1) pre-CS period, a 280 ms baseline period prior to the onset of the tone CS; (2) startle response (SR) period, first 80 ms after CS onset (EMG activity relating to a non-associative short-latency startle reaction elicited by the CS); (3) total conditioned response (CR) period, EMG activity that occurred during either the 200 ms (PRD280) or 800 ms (PRD880) of CS presentation that preceded onset of the US; (4) adaptive CR period, EMG activity that occurred during the 200-ms of CS presentation that preceded onset of the US (for PRD280 this is also the total CR period); (5) unconditioned response (UR) period, EMG activity that occurred from the programmed onset of the US to the end of the trial (240 ms) On paired trials, the EMG signal was shunted to during the US presentation to prevent intrusion of signal artifact from the electrical stimulation 594 Developmental Psychobiology DOI 10.1002/dev Tran, Stanton, and Goodlett Using criteria described by Skelton (1988) and Stanton et al (1992), any EMG response during the SR, CR (total/ adaptive), or UR periods that exceeded 0.4 arbitrary units above the pre-CS baseline mean was registered Note that measuring the adaptive CR limits the intrusion of spontaneous eyeblinks that can occur with a higher probability with the 880-ms ISI of the long-delay procedures Data Analysis Data from a total of 87 rats were considered for all analyses Mixed-design ANOVAs were used to analyze the following learning measures: Adaptive and total CR frequency (as percentage of trials) during paired CS–US trials (90 trials each session) and CR amplitude (V), latency to peak CR (ms), latency to CR onset (ms), CR area-under-the-curve (mm2, arbitrary units) during CS-alone trials (10 trials each session) During paired CS–US trials, the amplifier was gated for 100 ms during the UR period, preventing measurement of the full CR (240 ms additional time after US onset); analysis of CS-alone trials was therefore appropriate for CR latency measures Nonassociative and sensory measures (SR frequency and amplitude, and UR amplitude) were subjected to the same analyses, with exceptions as noted below For the dependent measures (frequency, amplitude, latency) of learning (CRs) and performance (URs and SRs), all initial analyses indicated that there were no main or interactive effects of sex, so the eyeblink conditioning data are presented with males and females combined and sex was not included as a factor in the ANOVAs These measures were analyzed with mixed ANOVAs with ISI training condition (2) and neonatal treatment group [‘‘Group’’] (3) as the between-subjects factors and session (6), as a repeated factor For CR frequency, the focus of the reported data was on adaptive CRs (those occurring 200 ms before the onset of the shock US) Comparable patterns of results were observed when percentage of CRs expressed during the entire CR collection period were analyzed, but those analyses are not reported to avoid redundancy Unconditioned response amplitudes during the final 80 paired CS–US trials of Session were analyzed to assess whether the US stimulation set individually for each rat (as previously described) produced comparable URs across groups Session was analyzed because the lower CR frequency in this session limited the extent to which summation effects of the CR and the UR EMGs would intrude on the measured UR Learning and performance measures were also followed up within each ISI training condition using Group  Session mixed ANOVAs; relevant interaction effects were subjected to simple effects tests and group main effects were identified with Tukey’s HSD post hoc tests For CR frequency and amplitude measures, post hoc tests were conducted on data averaged over Sessions and to evaluate terminal performance Furthermore, the pertinent dependent measures for evaluating the timing of conditioned eyeblink responses were latency to CR onset and latency to peak total CR during CS-alone trials Some subjects may not express any CR latency data during a given session, therefore all latency data were averaged across all six sessions and analyzed using between-subjects ANOVAs Data from rats that received paired training were analyzed separately from those that received explicitly unpaired training to assess the effects of neonatal treatment and ISI training condition on acquisition of short- and long-delay training For explicitly unpaired data, the measures analyzed included only CR frequency/amplitude and performance (SR frequency/amplitude and UR amplitude during the final 80 US-only trials of Session 1) Somatic growth (in paired-trained rats) as measured by body weight (g) during the six days of neonatal treatment (PD 4–9) was analyzed with a (Sex)  (Group)  Day (6) mixed ANOVA Body weights on PD 21 and PD 30 were analyzed with separate (Group)  (Sex) factorial ANOVAs Blood ethanol concentrations (BECs) were analyzed with a (ISI)  (Training, paired vs explicitly unpaired) between-subjects ANOVA Mean values were reported as mean Ỉ standard error of the mean (SE) and all significant results met a minimum alpha level of 05 Results Growth and BECs As we have consistently reported (Goodlett et al., 1998; Johnson & Goodlett, 2002; Tran et al., 2005), the ethanol intubations produced a modest but significant growth lag during the neonatal treatment period The ANOVA on body weights on PD 4–9 yielded the expected effects of group, F(2, 44) ¼ 19.24, Group  Day interaction, F(10, 220) ¼ 17.98, as well as day, F(5, 215) ¼ 1390.98; there were no other significant main or interactive effects The Group  Day interaction was mainly the result of lower body weights in EtOH-treated pups during PD 4–8 Tukey’s HSD post hoc analysis confirmed that the main effect of group was due to significantly lower body weights in all EtOH-treated pups (mean of PD 4–9) compared to SI and UC pups, which did not differ from each other By PD 21 and PD 30, body weight differences between EtOH-treated rats and controls were no longer significantly different, though by PD 30 the expected differences in body weight between males and females was evident, F(1, 44) ¼ 48.71 The mean BEC of all EtOH pups on PD was 424 Æ 11 mg/ dl The BECs were not different between EtOH pups as a function of ISI training condition or whether they received paired CS–US training or explicitly unpaired training (p’s > 20) The mean BEC for each EtOH group was: PRD280 EtOH ẳ 404 ặ 14; PRD880 EtOH ẳ 441 ặ 25; UNPRD380 EtOH ẳ 431 ặ 32; and UNPRD980 EtOH ẳ 418 ặ 19 Acquisition of Adaptive CRs With Paired Training: Frequency As shown in Figure 2A, the groups given paired CS–US presentations showed increases in CR frequencies over training sessions, whereas the groups given explicitly unpaired stimulus presentations showed no systematic change in CRs over sessions, confirming the associative control over CR acquisition The EtOH groups were impaired relative to both control groups and at both the short- and long-delay training ISIs (see Fig 2A) All groups showed increases in adaptive CRs over sessions [main effect of session, F(5, 250) ¼ 90.84], but the slower acquisition by the EtOH groups yielded a significant main effect of group, F(2, 50) ¼ 15.32, and a Group  Session interaction, F(10, 250) ¼ 2.57 The long-delay training produced slower rates of acquisition than short-delay training [ISI  Session interaction, F(5, 250) ¼ 2.64] The ISI  Session interaction was isolated by simple effects post hoc analyses on each session (pooled across treatment group) The groups given short-delay training showed Developmental Psychobiology DOI 10.1002/dev Ethanol-Induced Eyeblink Conditioning Deficits 595 FIGURE Mean (ỈSE) percentage of adaptive CRs (Panel A) and adaptive CR amplitudes (Panel B) for periadolescent rats trained on short-delay ISI (D280) or long-delay ISI (D880) procedures, using paired CS–US (PRD, filled lines) or unpaired stimulus presentations (UNPRD, dashed lines) for each of six training sessions (S1–S6) CR amplitude was measured in volt (V) units significantly fewer adaptive CRs than long-delay groups during Session only There were no other significant effects from the mixed ANOVA Follow-up mixed ANOVAs conducted separately on each ISI condition confirmed the group differences for the 280 ms ISI training [group: F(2, 26) ¼ 11.84, session: F(5, 130) ¼ 78.58, and Group  Session: F(10, 130) ¼ 2.36], and for the 880 ms ISI training [group: F(2, 24) ¼ 4.76, session: F(5, 120) ¼ 25.40, Group  Session non-significant] Terminal performance (Sessions and 6) within each ISI was analyzed using Tukey’s HSD post hoc tests These confirmed that for short-delay training, the EtOH group executed significantly fewer adaptive CRs during these terminal sessions than either control group, which did not differ from each other (EtOH ẳ 49 ặ 6%; SI ẳ 77 ặ 6%; UC ẳ 85 ặ 7%) Similarly for long-delay training, the terminal performance of the EtOH group was significantly lower than either control groups, which did not differ from each other (EtOH ẳ 42 ặ 7%; SI ẳ 74 ặ 7%; UC ẳ 76 ặ 9%) Acquisition of Adaptive CRs With Paired Training: Amplitude Analysis of the peak amplitude of adaptive CRs during the CS-alone trials was conducted in the same manner as for percentage of adaptive CRs The EtOH groups showed slower acquisition and smaller CR amplitudes than either the SI or the UC controls in both short-delay and long-delay training (see Fig 2B) The EtOH treatment group differences were confirmed by a significant main effect of group, F(2, 50) ¼ 11.63, ISI, F(1, 50) ¼ 4.13, and a significant Group  Session interaction, F(10, 250) ¼ 5.83 In addition, all groups showed increased CR amplitudes over sessions, main effect of session: F(5, 250) ¼ 57.1, and the increase in the CR amplitudes, as expected, was less for the long-delay groups than for the shortdelay groups [ISI  Session interaction, F(5, 250) ¼ 6.0] Post hoc analysis of the ISI  Session interaction showed that terminal CR amplitude (Sessions and 6) was significantly less for rats given long-delay training than those given short-delay training No other interactive effects were significant Follow-up mixed ANOVAs were conducted separately for each ISI training condition For the short-delay training, there were significant main effects of group, F(2, 26) ¼ 7.68, Session, F(5, 130) ¼ 101.22, and their interaction, F(10, 130) ¼ 7.63 Likewise, for the long-delay training, there were significant main effects of group, F(2, 24) ¼ 4.17, Session, F(5, 120) ¼ 16.84, and their interaction, F(10, 120) ¼ 2.55 Tukey’s HSD post hoc tests on the CR amplitudes during terminal performance also confirmed that for both ISI conditions, the EtOH group had significantly lower adaptive CR amplitudes in these terminal sessions than the SI and UC groups, which did not differ from 596 Developmental Psychobiology DOI 10.1002/dev Tran, Stanton, and Goodlett each other The group means (V) were as follows—PRD280: EtOH ¼ 2.0 ặ 0.7; SI ẳ 4.8 ặ 0.7; UC ẳ 6.2 ặ 0.8; PRD880: EtOH ẳ 1.1 ặ 0.5; SI ¼ 3.2 Ỉ 0.5; UC ¼ 3.7 Ỉ 0.6 The same pattern of results was observed for amplitude of CRs expressed during the entire CR collection period Timing of CRs: Latency to Peak CR and Latency to CR Onset Nine of the ethanol-treated rats (PRD280 ¼ 4, PRD880 ¼ 5) failed to exhibit reliable CR measurements for three or more sessions due to poor acquisition performance A (ISI)  (Group) ANOVA on CR peak latency confirmed the expected significantly longer peak latencies of the long-delay groups [main effect of ISI, F(1, 50) ¼ 355.16], but there were no other significant main or interactive effects The main effect of ISI was due to longer latencies of rats in the long-delay condition (757 Ỉ 17 ms) compared to those in the short-delay condition (305 Ỉ 17 ms) For CR onset latency, the two-way ANOVA again yielded the expected significant main effect of ISI, F(1, 50) ¼ 185.34, but no other significant main effects or interactions The main effect of ISI was due to longer onset latencies of rats in the long-delay condition (540 Ỉ 18 ms) compared to those in the short-delay condition (198 Ỉ 17 ms) Thus, despite the impaired acquisition of CRs by the EtOH groups, the timing of the CRs that were executed by the EtOH group (relative to CS onset and time of US presentation) was not different from controls A summary of the means for both latency measures is shown in Table Non-Associative Responding A summary of the mean UR amplitudes during Session can be found in Table A (ISI)  (Group) between-subjects ANOVA indicated no significant main or interactive effects of these factors This is consistent with our previous findings that the ability to produce shock-elicited unconditioned eyeblinks was not affected by neonatal ethanol treatment (Stanton et al., 1998; Tran et al., 2005) A (ISI)  (Group)  (Session) with session as the Table repeated factor on percentage and amplitude of SRs, indicated no significant main or interactive effects Startle responses were infrequent in all groups, with the percentage of trials containing an SR averaging (across six sessions) between 1.2 and 4.1%, regardless of ISI When they did occur, SR amplitudes were small (group means of 0.023, 0.039, and 0.064 V for EtOH, SI, and UC, respectively), but the differences between EtOH and UC did reach significance Explicitly Unpaired Stimulus Presentations For eyeblinks emitted during the CS presentations of the explicitly unpaired training, there were no significant main or interactive effects of Group or ISI condition The mean percentage of CRs never surpassed 10% for any group in either ISI conditions, and mean CR amplitude never surpassed 0.6 V for any group or ISI condition In addition, the SRs during the CS presentations were infrequent ( 40) The 598 Tran, Stanton, and Goodlett Developmental Psychobiology DOI 10.1002/dev FIGURE Mean (ỈSE) percentage of adaptive CRs (Panel A) and adaptive CR amplitudes (Panel B) for adult rats trained on short-delay ISI (D280) or long-delay ISI (D880) using paired CS–US (PRD, filled lines) or unpaired (UNPRD, dashed lines) procedures over the six training sessions (S1– S6) CR amplitude was measured in volt (V) units mean BEC for each EtOH group was: PRD280 EtOH ẳ 427 ặ 4; PRD880 EtOH ẳ 422 ặ 5; UNPRD380 EtOH ẳ 421 Æ 4; and UNPRD980 EtOH ¼ 394 Æ 27 Acquisition of Adaptive CRs With Paired Training: Frequency The neonatal alcohol treatment resulted in impaired acquisition of eyeblink CRs, regardless of ISI condition (see Fig 3A), as confirmed by a significant main effect of group, F(2, 49) ¼ 20.83 All groups given paired training showed increased CR frequencies over training [main effect of Session, F(5, 245) ¼ 96.57], but the longdelay training was acquired more slowly than the shortdelay training [main effect of ISI, F(1, 49) ¼ 8.49, and a significant ISI  Session interaction, F(5, 245) ¼ 3.02] Simple effects tests conducted on the ISI  Session interaction showed that groups given long-delay training showed significantly fewer adaptive CRs than short-delay groups during Sessions and 4, but not during the initial sessions (1 or 2) or terminal sessions (5 or 6) There were no other significant effects from the mixed ANOVA Follow-up (Group)  (Session) mixed ANOVAs conducted separately for each ISI condition confirmed the ethanol-induced learning deficits For the 280-ms ISI, there were significant main effects of group, F(2, 23) ¼ 18.41, and Session, F(5, 115) ¼ 57.0, but no interactive effects Tukey’s HSD post hoc tests confirmed significantly poorer terminal performance (Sessions and 6) of the EtOH group (53 Ỉ 7% compared to the SI (80 Ỉ 7%) and UC (90 Ỉ 8%) groups, which did not differ from each other For the 880-ms ISI, there were also significant main effects of group, F(2, 26) ¼ 5.94, and Session, F(5, 130) ¼ 40.14, but no interaction Tukey’s HSD again confirmed significant differences in mean terminal performance during long-delay training between the EtOH group (39 Æ 8%) and the SI (72 Æ 8%) and UC (68 Ỉ 8%) groups, which did not differ from each other A fully comparable pattern of results was observed for percentage of CRs expressed during the total CR collection period as for the adaptive CR period reported above Acquisition of Adaptive CRs With Paired Training: Amplitude The neonatal ethanol treatments also produced significant deficits in ECC conditioning as Developmental Psychobiology DOI 10.1002/dev measured by CR amplitude (see Fig 3B) A (ISI)  (Group)  (Session) mixed ANOVA indicated significant main effects of group, F(2, 49) ¼ 26.1, ISI, F(1, 49) ¼ 28.59, and session, F(5, 245) ¼ 57.14, and significant interactive effects of ISI  Group, F(2, 49) ¼ 5.15, ISI  Session, F(5, 245) ¼ 9.67, Group  Session, F(10, 245) ¼ 5.26, and ISI  Group  Session, F(10, 245) ¼ 2.53 Analyses of the ISI  Group  Session interaction indicated that the ethanol-treated rats had significantly lower CR amplitudes than controls regardless of ISI condition However, the UC and SI control groups generated significantly higher CR amplitudes during short-delay training than long-delay training (Sessions 3–6), whereas the CR peak amplitudes of the ethanol-treated rats given short-delay training were not different from peak CR amplitudes of ethanol-treated rats given long-delay training (all sessions) Follow-up mixed ANOVAs were conducted separately on each ISI condition to confirm the EtOH-induced deficits For short-delay training, there were significant main effects of group, F(2, 23) ¼ 15.67, and Session, F(5, 115) ¼ 34.09, and a significant interaction, F(10, 115) ¼ 4.23 Simple effects analyses indicated that ethanol-treated rats had lower CR amplitudes compared to UC rats during Sessions 2–6, and compared to SI rats during Sessions 4–6; SI and UC rats did not differ from each other significantly during any session For longdelay training, a similar pattern was evident, with significant main effects of group, F(2, 26) ¼ 9.97, Session, F(5, 130) ¼ 25.76, and Group  Session, F(10, 130) ¼ 2.32 The significant interaction was mainly due to lower CR amplitudes expressed by ethanol-treated rats compared to controls during the later sessions (Sessions 3–6 compared to UC rats; Session compared to SI rats); SI and UC rats not differ significantly from each other during any session A similar pattern of results was observed for CR amplitude during the total CR collection period (PRD280 ¼ 200 ms, PRD880 ¼ 800 ms) in paired CS–US trials as reported above for adaptive CR amplitude Timing of CRs: Latency to Peak CR and Latency to CR Onset A summary of the means for both latency measures is shown in Table The timing of CR peak latency and CR onset examined in the CS-alone trials in the same manner as described in Experiment (data averaged across six sessions) The between-subjects ANOVA confirmed the expected main effect of ISI, F(1, 48) ¼ 710.59, due to the later peak latencies of the longdelay groups; no other main or interactive effects were significant For CR onset latency, the ANOVA yielded both the significant main effect of ISI, F(1, 48) ¼ 229.38, and a significant main effect of treatment group, F(2, 48) ¼ 5.09; the interaction was not significant The main Ethanol-Induced Eyeblink Conditioning Deficits 599 effect of group was due mainly to the longer CR onset latency of the EtOH group relative to the UC controls in both training conditions; the SI and UC rats did not differ significantly from each other in either ISI training condition Non-Associative Responding UR amplitude data were analyzed as for Experiment (data averaged over the final 80 paired CS–US trials of Session 1) A summary of the mean UR amplitudes during Session is provided in Table A (ISI)  (Group) between-subjects ANOVA yielded no significant main or interactive effects of these factors The lack of neonatal treatment effects on UR amplitudes (and the lack of group differences in the US intensity required to support reliable conditioning) is consistent with our previous findings that shock-elicited unconditioned eyeblinks in adulthood are unaffected by neonatal ethanol treatment (Green et al., 2000) Startle responses (SR) during the first 80 ms after tone onset during paired training were again very infrequent in all groups, with the percentage of trials containing an SR averaging less than 5% in all groups The mixed ANOVA performed on percentage and amplitude of SRs yielded only a significant main effect of ISI, F(1, 49) ¼ 8.12; no other main or interactive effects were significant The effect of ISI was due to the modestly higher rate of SRs in the short-delay groups (4.2 Æ 0.5%) than in the long-delay groups (2.2 Æ 0.5%) Explicitly Unpaired Stimulus Presentations Eyeblinks emitted during the CS presentations of the explicitly unpaired training groups were analyzed using a (ISI)  (Group)  (Session) mixed ANOVA There were no significant differences among groups, between ISI conditions, over sessions, or any interaction of these factors Overall, mean percentage of CRs never surpassed 15% for any group across both ISI conditions, and mean CR amplitude never surpassed 0.6 V for any group across both ISI conditions In addition, the SRs during the CS presentations were infrequent (